Today’s jobs report made clear that despite rising interest rates and incessant recession talk, American businesses are still hiring. This was reflected mainly in the headline 315,000-job increase in nonfarm payroll employment, but also in the less-closely watched employment estimate from the household survey that is used to determine the unemployment rate, which showed an August gain of 442,000. Adjust the household survey number to measure something more or less equivalent to nonfarm payrolls, and the increase was 599,000.
For a while this spring, the household survey seemed to be showing hardly any employment growth even as nonfarm payrolls kept rising at a healthy pace. Now — if you make the proper adjustments — both measures are showing solid job growth. Here’s another way of looking at the data:
The Current Employment Survey that generates the nonfarm payrolls number is a mostly online check-in by the Bureau of Labor Statistics with “131,000 businesses and government agencies, representing approximately 670,000 individual worksites.” For the Current Population Survey, the Census Bureau quizzes “a probability selected sample of about 60,000 occupied households” by phone and in person about their employment situation and other matters.
As it says on the tin, the nonfarm payrolls number excludes farmers and farmworkers and those who aren’t on someone’s payroll — i.e. the self-employed. The CES also counts jobs rather than people, meaning that those with more than one job can show up multiple times in the payroll data while being counted as just one employed person in the household survey. Since March, agricultural employment and self-employment are down, while multiple job holding is up.
The BLS factors these three metrics into the “ CPS employment adjusted for CES concepts” data series shown in the first two charts, and they appear to account for about two-thirds of the difference since March between it and the unadjusted CPS employment numbers. I realize one has to squint to see the changes since March on the above chart, but figured that showing the numbers since the beginning of 2021 would give a sense of both how modest those changes have been and how common such ups and downs are in the household data. Other contributors to the disparity include workers in private households such as nannies and housekeepers, unpaid family workers in family-owned businesses and wage and salary workers with unpaid absences, for which the data I could find are not seasonally adjusted and so noisy that putting them on a chart seemed pointless.
The nonfarm payrolls numbers tend to be less noisy but are subject to revision; first as more information from employers trickles in during the next two months after the initial report, and then once a year in January when the numbers are trued-up with state and federal unemployment-insurance data for more than 10 million workplaces. The BLS gave a preliminary indication a few weeks ago of how big next January’s revision will be, reporting that nonfarm payroll employment as of March 2022 was 462,000 or 0.3% higher than the number as currently reported.
This provides no information on what’s happened since March and there’s a long history of economy-watchers claiming to see portents in the household survey that the establishment survey isn’t reflecting yet. That’s in part because it takes the BLS a while to make contact with newly formed businesses, and to figure out when businesses that aren’t reporting have shut down. To account for these comings and goings, the agency uses the ominous-sounding “ net birth-death model,” which generally does a good job but may be less reliable at economic turning points. With the numbers from both surveys headed in the same direction again, though, we do not seem to be at one of those turning points.
Bloomberg News provided this article. For more articles like this please visit
bloomberg.com.
Read more articles by Justin Fox
