by Bill McBride on 3/08/2010 03:43:00 PM
Monday, March 08, 2010
If you thought Snowmageddon distorted the employment figures in February, just wait until March. (ht JH)
The BLS could report a March headline number of 200,000 net payroll jobs, and that could be viewed as a weak report.
The March report will be distorted by two factors: 1) any bounce back from the snow storms, and 2) the decennial Census hiring that picked up sharply in March.
Many analysts considered the February BLS report - showing a headline net loss of 36,000 jobs - as an improvement over January because of snow factors. The general view is the snowstorms subtracted 50,000 to 150,000 payroll jobs from the report. I think this is a huge unknown, but I think the actual impact was fairly low.
Also in February, the Census added 15,000 workers. Although this number is NSA (Not Seasonally Adjusted), there really is no seasonal factor for this hiring. If we estimate job losses at 50,000 ex-Census in February, this would imply a range of 0 to 100,000 jobs would have been added without snow factors - if the analysts are correct about the impact of the snow storm.
Also the Census will add something like 100,000 workers to the March report. Luckily the Census Bureau reports the Census hiring - May will be the really crazy month with 100s of thousands of workers added to follow-up on anyone who didn't mail back the Census (Mail it back!).
Add the numbers up: If the economy added no jobs in February (after snow effects) and no jobs in March, the March employment number would be 150,000 (50,000 bounce back from the snow, 0 for March, and 100,000 Census hiring).
If the economy added 50,000 jobs in February (after snow) - the consensus view - and adds another 50,000 in March, the March number will be around 250,000. And remember it takes 100,000 to 150,000 net jobs added per month to stay up with population growth. So a headline number of 250,000 - though an improvement - would still be a weak report because it just means 50,000 in March after distortions.
And if the economy added 100,000 jobs in February (very unlikely in my view), and another 100,000 in March - the bounce back would be 150,000 + 100,000 for March + 100,000 for the Census - and we would see a report of 350,000 jobs!
The Census distortions will last most of the year. The Census will add jobs through May, and then subtract jobs for the following 6+ months. Right now it is making the payroll report look better and will lower the unemployment rate slightly over the next few months (just 0.1% to 0.2%), and starting in June, Census hiring will make the payroll report look worse - but the net will be close to zero by the end of the year.