Tuesday, June 01, 2010

Impact of Census 2010 on Payroll Report

by Bill McBride on 6/01/2010 02:49:00 PM

We are starting to see articles like this from CNBC: Strong Jobs Number on Friday Could Give the Markets a Boost

Economists expect the US economy generated about 540,000 jobs in May—a large portion of which expected to come from Census hiring—and many analysts will be hoping that's enough to assuage investor fears that the European debt contagion could cause a double-dip recession.
The BLS will release the May employment report on Friday. The consensus is for a gain of 540,000 payroll jobs in May, and for the unemployment rate to decline slightly to 9.8% (from 9.9%).

As the CNBC article noted, a large portion of the payroll jobs in May will be temporary hires for Census 2010 (May is the peak month). It will be important to remove the Census hiring to try to determine the underlying trend.

We can estimate the Census hiring using weekly payroll data from the Census bureau (ht Bob_in_MA). If we subtract the number of Temporary 2010 Census Workers in the 2nd week of May from the number in the second week of April, this suggests the Census boost will be around 417K in May. The Census Bureau will release the actual number with the employment report.

Impact Census Hiring Click on graph for larger image in new window.

This graph shows the actual impact of Census hiring in 1990, 2000, and through April 2010. The impact of the Census hiring, from May through December 2010, are my preliminary estimates.

When the employment report is released on Friday, a key number will be payroll jobs ex-Census - since the Census will probably add over 400,000 temporary payroll jobs (these are real jobs, but they mask the underlying trend). This temporary hiring will also push down the unemployment rate in May by 0.1% or 0.2% based on previous decennial Census hiring.

Most ex-Census forecasts are in the 130,000 to 150,000 range (although most forecasts only release the headline number). This would be a decrease from the 224,000 ex-Census payroll jobs in April.

Starting in June, the Census will negatively impact the payroll report. My preliminary estimate is for a decline of 200,000 Census payroll jobs in June (see graph above). If the underlying trend is a positive 200,000 payroll jobs, the headline number will be zero! And that will understate the underlying trend, just like the 500,000+ will overstate the trend in May. So we will need to adjust for the decennial Census for most of this year.