by Calculated Risk on 6/07/2010 12:00:00 PM

Last week I posted the Impact of Census 2010 on Payroll Report

My estimate was that the 2010 Census would add 417,000 payroll jobs in May; the actual was 411,000 payroll jobs.

My preliminary estimate is the Census will subtract 200,000 payroll jobs in June - and most of the remaining temporary Census jobs (564,000 total in May) will be unwound by September.

I've been puzzling over how much (if any) these temporary jobs lowered the unemployment rate in May. I think these workers come from three groups:
1) already employed workers taking a part time job,
2) people not in the workforce picking up a little temporary income (like retirees or students who would otherwise not be in the workforce), and
3) the unemployed taking a part time job.

Sure enough there was an increase in people working multiple jobs in May. The number of multiple jobholders jumped by 210,000 in May (seasonally adjusted). There can be other reasons for this increase, but if we assume these are mostly Census workers, then about half the 411,000 additional Census workers already have other jobs - so for these workers, the temporary Census jobs has no impact on the unemployment rate.

The other half are probably otherwise unemployed workers, or people not in the workforce (although the participation rate declined in May). If we assume that this is mostly unemployed workers, these temporary hires lowered the unemployment rate by around 0.1% (from 9.8% to 9.7%).