Private employers added 86,000 jobs in October, according to ADP's November 4 report. September's gain was revised down from 90,000 to 86,000, leaving the two months level on the latest estimates.
For people looking for work, the detail is uneven. Trade, transportation and utilities added 39,000 jobs; education and health services added 28,000. Professional and business services lost 27,000, its third consecutive monthly decline. An expanding national payroll can coexist with a difficult search in a particular industry.
The next federal employment report is due Friday. It will answer questions that Wednesday's private payroll count cannot, including what happened to unemployment across the country.
ADP also reported median base pay growth of 3.2% over the year and gross pay growth of 4.7%. For job-changers, the respective figures were 4.8% and 7.6%. ADP attributed the larger earnings gain to a combination of base pay, bonuses, overtime and additional hours. A higher annual earnings figure can therefore include more hours worked as well as a higher rate of pay. Our guide to wages and purchasing power explains the separate effects of inflation and benefits.
What the 86,000 figure measures
ADP draws on anonymized payroll records covering more than 26 million private-sector employees. Its headline measures the change in employment. Government jobs fall outside that private-sector count.
The difference between new hiring and net job growth becomes clearer alongside the federal Job Openings and Labor Turnover Survey. That report, released November 3, concerns September, a month earlier than the new ADP estimate.
BLS estimated 5.4 million hires and 5.3 million separations in September across the nonfarm economy. Separations include people quitting, being laid off or discharged, retiring and leaving for other reasons. Millions of people can start jobs during a month even when the overall number of jobs changes much less.
Those rounded totals cover a different survey and period from October's ADP figure. Subtracting them does not reproduce ADP's estimate. Net growth measures the balance after jobs are added and lost.
Openings are a separate count
September had an estimated 7.5 million job openings, little changed over the month, according to BLS. The agency counts positions open on the last business day of the month. Hires cover additions to payroll throughout the month.
An opening therefore does not establish that a person was hired during that period. Nor should the openings total be compared directly with a monthly net payroll increase as if both counted the same activity.
The survey also estimated 3.4 million quits and 1.6 million layoffs and discharges in September. Quits increased; layoffs and discharges changed little. These measures help distinguish voluntary departures from employers ending jobs, a distinction that a single net employment number cannot provide.
Friday brings two views of employment
The BLS release calendar lists October's Employment Situation for November 6 at 8:30 a.m. Eastern, or 5:30 a.m. Pacific.
That release combines two surveys. The establishment survey measures nonfarm payroll jobs, hours and earnings using information from employers, including government agencies. The household survey supplies the unemployment rate and counts employed people, with broader coverage that includes the self-employed and agricultural workers.
The agency's survey comparison explains a further difference: a person with multiple nonfarm payroll jobs can be counted for each job in the establishment survey, while the household survey counts that employed person once. Different definitions and methods can produce different monthly changes.
For that reason, ADP's private payroll estimate cannot settle what Friday's unemployment rate will be. The federal payroll total also covers a wider set of employers. Comparing the private-sector components is closer in scope than comparing ADP directly with the entire nonfarm payroll headline.
The BLS payroll survey uses the employer's pay period containing the 12th of the month. Its household survey generally uses the calendar week containing that date. The periods can differ even within the same monthly release.
