Guide · Markets
Nonfarm Payrolls: How to Read the US Jobs Report
Nonfarm payrolls count how many jobs US employers added in a month, and the number moves bonds, the dollar and stocks within seconds. Here is how to read the headline, the revisions, the unemployment rate and wages, using September 2026's +29,000 report.
At 8:30 a.m. Eastern time on Friday, October 2, 2026, the Bureau of Labor Statistics reported that US employers had added 29,000 jobs in September. Economists had expected about 90,000, and the figure was below the forecast of every economics team surveyed by Bloomberg, according to First Trust. Worse, payroll gains for July and August were revised down by a combined 60,000, so the net change including revisions was a loss of 31,000 jobs. The unemployment rate rose to 4.2% from 4.1%. The next report, for October, is due on Friday, November 6, with a FactSet consensus of +55,000. Before that number arrives, it helps to know which parts of the release markets actually trade.
The release, officially called the Employment Situation, combines two separate surveys. The establishment survey asks around 120,000 businesses and government agencies how many people are on their payrolls; that is where the headline nonfarm payrolls number, the industry breakdown, average hourly earnings and the average workweek come from. Farm workers, the self-employed and household staff are excluded, hence "nonfarm". The household survey, the Current Population Survey, interviews about 60,000 households and produces the unemployment rate, the participation rate and a second count of employment that does include the self-employed.
The two surveys often disagree, and September 2026 was a textbook case. While payrolls rose 29,000, household employment rose 406,000 and the labour force grew by 485,000, lifting the participation rate to 61.8% from 61.6%. More people looking for work is also why unemployment can rise in a month when employment rises: the unrounded rate went from 4.141% to 4.175%, First Trust calculated, so the 0.1-point jump was largely a rounding effect. The household survey is much noisier month to month because its sample is smaller, which is why markets lean on payrolls, but a gap that persists for several months is worth noticing.
Revisions are the part beginners skip and professionals read first. Each report revises the two previous months as more employers respond, and once a year the BLS benchmarks the whole series to state unemployment-insurance records. In a slowing economy revisions tend to be negative, because late responders are more often struggling businesses, so a soft headline with large downward revisions is a weaker report than the headline alone. September's private payrolls rose 46,000 but prior months were cut by 81,000. A useful habit is to track the three-month average including revisions rather than any single month.
Then look at the mix and the pay. In September health care and social assistance added 23,000 jobs, construction 11,000 and manufacturing 9,000, while information lost 10,000 and government 17,000; federal payrolls were down 328,000 since January 2025. Average hourly earnings rose 0.1% on the month and 3.0% on the year, slower than consumer prices, which tells the Federal Reserve that wages are not fuelling inflation. For a central bank weighing a hike, as the Fed is with its range at 3.75% to 4%, a weak job count with soft wages argues for waiting; a strong count with hot wages argues for moving.
Markets react in a specific order. Two-year Treasury yields and fed funds futures move first, because they price the next rate decision; the dollar follows the yields; stocks are last and the most ambiguous, since good jobs news can be bad news for rates. To prepare for November 6, write down the consensus, the whisper range and what a surprise would mean for the December Fed meeting. Then, at 8:30, read in this order: headline versus consensus, revisions, unemployment rate, hourly earnings. A miss of 50,000 in either direction is within the survey's normal margin of error, about 136,000 according to the BLS, so the trend matters more than any one Friday.