U.S. Payrolls Add 162,000 in August 2026, Unemployment Holds at 4.1%

Total nonfarm payroll employment rose by 162,000 in August 2026 and the unemployment rate was unchanged at 4.1 percent, the Bureau of Labor Statistics reported. The gain carries information on its own: a monthly change of this size sits outside the establishment survey's sampling noise, while the 31,000 average monthly gain of the prior twelve months sits comfortably inside it. Employment rose in food services and drinking places and in local government education, and the information industry lost jobs.

Revisions Turn the Prior Two Months Around

BLS revised June up by 11,000, from a gain of 20,000 to a gain of 31,000, and July up by 44,000, from a loss of 23,000 to a gain of 21,000. With those changes, employment in June and July combined is 55,000 higher than previously reported. A net revision of that size is past the threshold at which revisions move markets, and both months moved in the same direction.

The July revision is the one that matters. As first published, July was a 23,000 decline — a labor market shedding jobs. It now reads as a 21,000 gain. That is a sign flip, not a magnitude adjustment, and it changes the shape of the summer from stall to slow-but-positive.

This is the Current Employment Statistics model behaving the way it does around turning points. The model leans on historical trends to estimate business formations and closures, which makes it a dependable estimator in steady conditions and a lagging one when conditions change. When the economy shifts velocity, initial prints tend to carry the old trend forward and the revisions do the correcting — in either direction. A run of downward revisions is what a decelerating labor market looks like inside that estimator; an upward net revision of this size is what the reverse looks like. Neither is evidence of manipulation, and both argue for treating any single initial print as provisional.

One caveat survives the good news: even after being revised up, neither June nor July is a month whose gain can be distinguished from zero at the establishment survey's precision. August is.

Where the Gains Came From: Sector Composition

The gains were broad in direction but concentrated in level. Leisure and hospitality led the month, with food services and drinking places accounting for essentially all of it. Government came next, carried by local government education, which added back most of what it surrendered the prior month. Education and health services placed third.

That composition is the soft spot in an otherwise firm report. Leisure and hospitality, education and health, and government are the three supersectors least informative about private cyclical demand — they grow on demographics, public budgets, and post-pandemic normalization largely regardless of where the business cycle sits. A print carried by those three can mask weak underlying private hiring.

The cyclical read is better than that framing implies. Manufacturing continued its upward trend, adding 16,000 and extending a run that has lifted the sector by 58,000 since a recent low in December 2025, with machinery and fabricated metal products both contributing. Construction added 22,000, with nonresidential specialty trade contractors continuing to trend up. Manufacturing and construction are the cyclically informative supersectors in this report, and both grew.

Information was the notable decline, losing 23,000 against an average of 8,000 per month over the prior twelve months. The declines were spread across computing infrastructure and data processing, publishing, and broadcasting and content providers — a technology and media contraction rather than a cyclical one.

Industry Detail Beneath the Supersectors

Health care added 13,000, well below its 32,000 average monthly gain over the prior twelve months, with home health care services and hospitals leading what growth there was. Health care has been a reliable source of payroll growth through this expansion, and a month well under its trailing average is the quiet warning in this report: the cushion that has been offsetting weakness elsewhere is thinner than it was.

Local government payrolls rose, with the increase concentrated in education after a decrease the prior month. Local government education has shown little net change since January 2025, so this is a give-back rather than a trend. Wholesale trade, retail trade, and transportation and warehousing all showed little change over the month, as did mining, professional and business services, social assistance, and other services.

Wage Growth Stays Inside the Non-Inflationary Range

Average hourly earnings for all private employees rose by 10 cents, or 0.3 percent, to $37.75 in August, and are up 3.1 percent over the year. Production and nonsupervisory earnings rose 11 cents, also 0.3 percent, to $32.53.

At 3.1 percent, annual wage growth sits inside the range consistent with the Federal Reserve's inflation target once ordinary productivity growth is accounted for. This is not a wage-driven inflation problem, and it has not been one for some time.

Hours moved up, which is the detail worth flagging:

• All private employees: the average workweek edged up 0.1 hour to 34.4 hours.
• Manufacturing: the average workweek edged up 0.1 hour to 40.5 hours, with overtime unchanged at 3.1 hours.
• Production and nonsupervisory employees: the average workweek remained at 33.8 hours.

Employers cut hours before they cut headcount, and they add hours before they add headcount. A workweek that lengthens alongside a payroll gain outside the survey's noise band is a more coherent signal than either would be on its own.

The Household Survey: Unemployment Holds as Participation Rises

The unemployment rate was unchanged at 4.1 percent, with 7.0 million people unemployed, and is down 0.2 percentage points over the year. The labor force participation rate edged up 0.2 percentage points to 61.6 percent, and the employment-population ratio was 59.1 percent.

That combination is the reverse of the one that flattered the prior report. An unemployment rate that falls while participation falls is arithmetic, not improvement — people who leave the labor force stop being counted as unemployed. Here the rate held while participation rose, meaning the labor force absorbed additional entrants without pushing the jobless rate up. This month the household survey corroborates the establishment survey instead of contradicting it.

The caveat is the trend behind the monthly move. Participation is still down 0.5 percentage points since January and 0.7 percentage points over the year. One month of recovery does not reverse that, and the August uptick has to repeat before it counts as a turn.

Underneath the headline rate, the slack measures were mixed:

• Part time for economic reasons fell by 414,000 to 4.4 million — people who wanted full-time work but were working part time because their hours were cut or they could not find full-time jobs.
• Long-term unemployed (jobless for 27 weeks or more) changed little at 1.9 million, and accounted for 27.0 percent of all unemployed people.
• Not in the labor force but wanting a job changed little at 5.7 million, of whom 1.7 million were marginally attached to the labor force and 441,000 were discouraged workers.

The drop in involuntary part-time work is a genuine improvement in job quality. The long-term share is the offsetting concern: more than a quarter of the unemployed have now been out of work for 27 weeks or more, a composition that does not improve simply because monthly hiring picks up.

Among the major worker groups, the jobless rate for people who are Asian declined to 3.2 percent, and the rate for teenagers edged up to 14.1 percent, mostly offsetting a decline in the prior month. Rates for the other major groups showed little change.

What to Watch

The Employment Situation for September 2026 publishes on Friday, October 2, 2026. Two data points will settle whether August was a turn or a bounce.

First, participation. If it holds at or above 61.6 percent while the unemployment rate stays near 4.1 percent, August's improvement was real absorption of labor supply and the household survey has genuinely turned. If participation falls back and the unemployment rate falls with it, August was noise around a still-shrinking labor force.

Second, what September's report does to August itself. If it leaves August near 162,000, the establishment survey caught the acceleration in real time and the pickup is in the data. If August is marked down toward the 31,000 twelve-month average, the model overshot in the direction opposite to its recent misses, and the labor market is roughly where it was in July.