Nonfarm payrolls rose more than expected last month, despite being held back by the strike at General Motors (which subtracted 42,000) and the exit of 20,000 temporary workers for the 2020 census. There is some uncertainty in these data. The birth/death model, which attempts to account for the impact of business creation and destruction, added 274,000 to the unadjusted payroll figure (about the same as in October 2018). The birth/death model does a good job, except at turning points. This is probably not a turning point, but you won’t know until after the fact. Demographic changes (an aging population and slower growth in the workforce) imply that payroll growth ought to be slowing down. However, it’s unclear how much slack is still available in the labor market – a key question for Fed policymakers.
Nonfarm payrolls rose by 128,000 in the initial estimate for October, with a net +95,000 revision to figures for August and September. The GM strike subtracted 42,000 (these workers will return to the nonfarm payroll total in November). Preliminary work for the 2020 census led to a 25,000 increase in government workers in August. With much of that work done, most of these jobs disappeared in October. The census will play a big part in the payroll figures next year. In the last two censuses, temporary hiring rose to over half a million by May of the census year and these jobs were shed in June, July, and August.
The establishment survey (which yields payrolls, hours, and average hourly earnings) samples about 142,000 businesses and government agencies. This results in some sampling error in the payroll data (the monthly change in payrolls is reported accurate to ±110,000). In addition, a number of businesses are created and ended each month (some seasonal in nature), and there is job creation and destruction associated with that. According to the BLS, “both the business birth and death portions of total employment are generally significant, but the net contribution is relatively small and stable.” The birth/death model does a yeoman’s job, except at turning points. Job losses during the financial crisis were much more severe in revision than initially reported. The Fed is well aware of this problem and has made some effort to better anticipate job declines. The birth/death model added 274,000 to the unadjusted payroll figure for October (vs. +275,000 in October 2018). That may have exaggerated the October gain in adjusted payrolls. Other labor market data suggest that this is not a turning point, but could be a soft patch).