by Calculated Risk on 10/05/2011 01:29:00 PM

On Friday the BLS will release the September Employment Situation Summary at 8:30 AM ET. Bloomberg is showing the consensus is for an increase of 65,000 payroll jobs in September, and for the unemployment rate to increase to 9.2% from 9.1% in August.

Overall the economic data for September was fairly weak, though mostly better than in August. The BLS reported zero jobs added in August, so "better" doesn't mean much. Of course, the Verizon labor dispute subtracted 45,000 payroll jobs in August, and these jobs will be added back in the September report. So better means more than 45,000.

Here is a summary of recent data:

• The ADP employment report showed an increase of 91,000 private sector payroll jobs in September. Unfortunately ADP hasn't been very useful in predicting the BLS report. Also note that government payrolls have been shrinking by about 35,000 on average per month this year. Also the ADP doesn't include the Verizon labor dispute, so this suggests around 91,000 private nonfarm payroll jobs added, plus 45,000 from Verizon, minus 35,000 government workers - or around 101,000 total jobs added in September.

• The ISM manufacturing employment index increased to 53.8% from 51.8% in August. Based on a historical correlation between the ISM index and the BLS employment report for manufacturing, this reading suggests an increase of 2,000 private payroll jobs for manufacturing in September. This is consistent with the regional Fed manufacturing surveys showing some hiring in September.

However, the ISM non-manufacturing employment index decreased 2.9 percentage points to 48.7 percent. According to ISM this means contraction, but a historical correlation suggests this indicates some hiring.

Click on graph for larger image in graph gallery.

This scatter graph compares the ISM non-manufacturing employment index (x-axis) and the BLS report for private service employment (as a percent change per month). This suggests private service employment growth of around 27,000 in September (not including Verizon).

So the ISM surveys suggests 27,000 service jobs added, 2,000 manufacturing jobs added, plus the 45,000 Verizon workers, minus 35,000 government jobs - or about 39,000 jobs added in September.

• Initial weekly unemployment claims averaged about 417,000 per week in September, up from the 411,000 average in August.

However, claims were elevated during the BLS reference week (includes the 12th of the month). For the middle two weeks of September, claims averaged 430,000 per week - suggesting some increase in layoffs mid-month.

• The final September Reuters / University of Michigan consumer sentiment index increased to 59.4 from 55.7 in August. This is frequently coincident with changes in the labor market, but also strongly related to gasoline prices and other factors. This was probably impacted by the debt ceiling debate (it usually take 2 to 4 months from sentiment to recover from an "event"). In in general this low level would suggest a weak labor market.

• And on the unemployment rate from Gallup: Gallup Finds U.S. Underemployment Stuck at 18.5% in Mid-Sept.

Unemployment, as measured by Gallup without seasonal adjustment, is 8.8% in mid-September -- down from 9.1% at the end of August and the same as it was at the end of July. However, the apparent improvement in unemployment from August to mid-September may merely reflect normal seasonal hiring patterns and not be an indication that the employment situation is improving. On the other hand, current unemployment is considerably better than the 9.4% of a year ago.

NOTE: The Gallup poll results are Not Seasonally Adjusted (NSA), so use with caution. Usually the NSA unemployment rate declines in September, so this would suggest little change in the unemployment rate from August.

There always seems to be some randomness to the employment report, but my guess is it will be in the 40,000 to 100,000 range. There were some clear negatives this month - weekly claims were elevated during the reference period, consumer sentiment was very low, and the ISM non-manufacturing employment index declined below 50. Still I'll take the over on the consensus (above 65,000), mostly because of the 45,000 jobs added from Verizon. Caveat: my track record when I take the under has been very good - but recently I've been mostly wrong when I've taken the over!