Nonfarm Payrolls +29K vs 90K Est — Unemployment 4.2%, Wages +0.1%, 60K Revised Away

Fundamentals · 2026-10-02

Total nonfarm payrolls +29,000 in September (vs 90,000 est, prior +133,000 revised from +162,000); July revised to -10,000 from +21,000, so July and August combined are 60,000 lower; private +46,000 vs 85,000 est, government -17,000; unemployment 4.2% vs 4.1% est (unrounded 4.18% from 4.14%); average hourly earnings +0.1% to $37.81 vs 0.3% est, 3.0% y/y vs 3.2% est; average workweek 34.4 hours vs 34.3 est; health care and social assistance +23,000, half of all private hiring; private diffusion index 49.0 from 57.6, more industries cutting than adding; temporary help -10,900; information -10,000; three-month average +51,000 against a 12-month average of 45,000; next release Friday 6 November.

What It Changes

• A miss on all three numbers the market trades. Payrolls missed by 61,000, unemployment rounded up to 4.2% and wages rose 0.1% against 0.3%. Every one of them landed on the soft side of what this morning's Open called the strong case.
• The revisions hurt more than the headline. July went from +21,000 to -10,000 and August from +162,000 to +133,000. Inside those totals, private payrolls for July and August were cut by 81,000; government revisions added back 21,000.
• For the rate path, it takes the labour market out of the argument for an October hike. The Fed hiked to 3.75%-4.00% on 16 September on inflation, not jobs, and October odds were already only about one in four before 08:30. A 0.1% wage print gives the hawks nothing on wage-driven inflation.
• It is a soft labour market, not a breaking one. The rise in unemployment came from people entering the labour force to look for work. Job losers fell 45,000, and claims printed 197,000 yesterday. Firms are not firing. They are barely hiring.

Impact

• USD — Bearish — the US half of the dollar's case was rate differentials, and all three headline numbers came in on the dovish side.
• Payrolls at 29,000 do not clear BLS's own 122,000 significance threshold. But a 61,000 miss, 60,000 of downward revisions and a 0.1% wage print point the same way.
• Wages at 3.0% y/y are now below August's 3.4% headline PCE, so real hourly pay is falling again. That weakens any wage-push argument for further hikes.
• The offset is Europe. The dollar sat near an 18-month high this morning because of French fiscal stress, not US strength, and this release does nothing about France.
• US Indices (ES / NQ / YM / RTY) — Mixed, lean bullish — in a hiking cycle, soft jobs and soft wages lower the odds of another hike, and that matters more to equities than a 29,000 headline.
• The read is strongest for RTY. Small caps carry the most floating-rate debt, so a lower front-end yield helps them most.
• NQ is next, as the long-duration index. YM is the weakest beneficiary because cyclical earnings rely more on hiring and hours.
• The cost: aggregate weekly hours were flat and aggregate weekly payrolls rose only 0.2%, against 0.7% in August. Labour income growth slowed sharply into the fourth quarter.
• Gold (GC) — Bullish — a softer labour print lowers the expected policy path, which lowers real yields and weakens the dollar, the two things gold has traded against all month.
• Gold was still -2.72% over five days before the release, per this morning's Open, so it starts from a weak position.
• If the October hike is priced out entirely, the remaining support depends on whether December's roughly 80% odds also fall.
• 6J — Bullish yen — this is the one pair with its own reason: lower US yields line up with a hot Tokyo core CPI (2.7%) and a possible BOJ hike on 29-30 October.

Inside The Number

Total nonfarm payrolls rose 29,000 in September against a consensus of 90,000 (Trading Economics; MarketWatch had 84,000). BLS's own summary: payrolls and unemployment "changed little," and "employment in all major industries changed little over the month." That is accurate. It is also the problem: no major industry grew enough to register.

The revisions are the most important part of the release. A month ago, August's 162,000 was the year's largest upside surprise, and it turned July's reported job loss into a gain. Both have now gone back. July is -10,000, a job loss again, and August is 133,000. The private revisions are bigger than the totals show. July private went from +71,000 to +28,000, and August private from +127,000 to +89,000, so 81,000 private jobs were removed from those two months. Government was revised up by 21,000. The three-month average is 51,000, private 54,000.

Health care and social assistance did half the private hiring. It added 23,000 of the 46,000 private jobs. Health care rose 17,000 (ambulatory +13,400, hospitals +12,000, nursing and residential care -8,700), below its 12-month average of 33,000. Remove the sector and total payrolls rose about 6,000. The private diffusion index fell to 49.0 from 57.6. Below 50, more of the 250 industries cut jobs than added them. The manufacturing diffusion index fell further, to 46.5 from 63.9.

Temporary help fell 10,900, the largest drop in employment services, which fell 14,300. Temporary staffing is usually the first place firms cut, and it has now fallen in two of the last three months. Professional and business services lost 9,000, information lost 10,000 after -18,000 in August, and financial activities lost 7,000. BLS notes financial activities is now down 129,000 since May 2025, with 90,000 of that in insurance carriers.

The household survey explains the 4.2%. Unemployment rose 78,000 to 7.11 million and the rate rounded up from 4.14% to 4.18%. Look at who the new unemployed are. Re-entrants rose 152,000 and new entrants 116,000, while job losers fell 45,000 and job leavers fell 173,000. Participation rose to 61.8% from 61.6%, and the labour force grew 485,000. People are coming back to look for work faster than firms are hiring them. That is a different problem from layoffs, and it matches claims at 197,000 and continuing claims at a 54-week low yesterday.

Wages are the dovish surprise. Average hourly earnings rose 5 cents, or 0.13%, to $37.81, and 3.0% over 12 months. Production and nonsupervisory pay rose 0.2% to $32.60. Hours held at 34.4, a tenth above consensus. With hours flat and pay barely up, aggregate weekly payrolls rose 0.2%, against 0.7% in August. Earlier this week, August PCE had headline inflation at 3.4% and core at 3.0%. Hourly pay is now growing below headline inflation and only in line with core, so labour costs are not where the inflation is coming from. ISM's prices index jumped to 77.9 yesterday. The pressure is in inputs, not wages.

ADP overstated private hiring this month, the opposite of August. On Wednesday ADP had private payrolls at +90,000; BLS has +46,000. In August the miss went the other way: ADP had +38,000 against BLS's first estimate of +127,000, which has since been revised to 89,000. On the detail, ADP had manufacturing +17,000 (BLS +9,000) and education and health +55,000 (BLS +20,000). The gap between the two series is wide and changes sign from month to month. It is not a reliable guide to the BLS print.

The Internals

The headline, consensus and revisions:

Measure · Sep 2026 · Aug 2026 · Jul 2026 · Consensus

Total nonfarm payrolls · +29,000 · +133,000, from +162,000 · -10,000, from +21,000 · +90,000

Private payrolls · +46,000 · +89,000, from +127,000 · +28,000, from +71,000 · +85,000

Government payrolls · -17,000 · +44,000, from +35,000 · -38,000, from -50,000 · n/a

Manufacturing payrolls · +9,000 · +15,000 · +20,000 · +10,000

Unemployment rate · 4.2% · 4.1% · 4.1% · 4.1%

Unemployment rate, unrounded · 4.18% · 4.14% · 4.09% · n/a

Average hourly earnings, m/m · +0.1% · +0.3% · n/a · +0.3%

Average hourly earnings, y/y · +3.0% · +3.1% · n/a · +3.2%

Average weekly hours · 34.4 · 34.4 · 34.3 · 34.3

Unrounded unemployment rates are computed from the household survey levels. The August y/y earnings figure is as published in August.

Trend and significance:

Measure · Reading

September payroll gain · +29,000

Three-month average, total nonfarm · +51,000

Three-month average, total private · +54,000

Average monthly gain, prior 12 months · +45,000

July and August combined revision · -60,000

July and August private revision · -81,000

BLS threshold for a significant monthly change · About 122,000

Does this print clear that threshold · No

Employment change by industry, over the month, in thousands:

Industry · Sep 2026 · Aug 2026 · Jul 2026 · Sep 2025

Total nonfarm · 29 · 133 · -10 · 76

Total private · 46 · 89 · 28 · 68

Goods-producing · 18 · 34 · 35 · 11

Mining and logging · -2 · 3 · -3 · -2

Construction · 11 · 16 · 18 · 16

Manufacturing · 9 · 15 · 20 · -3

Durable goods · 8 · 14 · 24 · -5

Motor vehicles and parts · 1.6 · -7.3 · 11.6 · -2.3

Nondurable goods · 1 · 1 · -4 · 2

Private service-providing · 28 · 55 · -7 · 57

Wholesale trade · 5.0 · 8.0 · 6.3 · 8.8

Retail trade · 5.8 · -5.1 · 5.6 · 13.1

Transportation and warehousing · 7.6 · 3.3 · 10.3 · -39.3

Utilities · 0.5 · 2.1 · 1.1 · 0.0

Information · -10 · -18 · 4 · -1

Financial activities · -7 · -9 · -10 · 4

Professional and business services · -9 · -5 · -6 · -16

Temporary help services · -10.9 · 0.2 · -10.0 · -11.7

Private education and health services · 20 · 38 · 14 · 54

Health care and social assistance · 23.0 · 35.1 · 13.3 · 55.3

Leisure and hospitality · 10 · 37 · -38 · 37

Other services · 6 · 3 · 6 · -3

Government · -17 · 44 · -38 · 8

Hours and earnings, all private employees:

Measure · Sep 2026 · Aug 2026 · Jul 2026 · Sep 2025

Average hourly earnings · $37.81 · $37.76 · $37.64 · $36.70

Average weekly hours · 34.4 · 34.4 · 34.3 · 34.2

Average weekly earnings · $1,300.66 · $1,298.94 · $1,291.05 · $1,255.14

Aggregate weekly hours, m/m · 0.0% · +0.4% · 0.0% · +0.1%

Aggregate weekly payrolls, m/m · +0.2% · +0.7% · +0.1% · +0.3%

Manufacturing workweek · 40.6 · 40.6 · 40.4 · 40.0

Manufacturing overtime · 3.0 · 3.0 · 3.1 · 2.9

Retail trade workweek · 30.6 · 30.2 · 30.1 · 29.9

Production and nonsupervisory hourly earnings · $32.60, +0.2% · n/a · n/a · n/a

Diffusion indexes, the share of industries adding jobs, where 50 is an even balance:

Index · Sep 2026 · Aug 2026 · Jul 2026 · Sep 2025

Total private, 250 industries · 49.0 · 57.6 · 54.4 · 50.4

Manufacturing, 72 industries · 46.5 · 63.9 · 52.1 · 47.2

Household survey, seasonally adjusted, in thousands:

Measure · Sep 2026 · Aug 2026 · Change · Sep 2025

Civilian labour force · 170,262 · 169,777 · +485 · 171,261

Employed · 163,152 · 162,746 · +406 · 163,656

Unemployed · 7,109 · 7,031 · +78 · 7,605

Not in labour force · 105,292 · 105,638 · -346 · 102,964

Participation rate · 61.8% · 61.6% · +0.2pp · 62.5%

Employment-population ratio · 59.2% · 59.1% · +0.1pp · 59.7%

U-6 underemployment rate · 7.6% · 7.7% · -0.1pp · n/a

Part time for economic reasons · 4,501 · 4,390 · +111 · 4,594

Slack work or business conditions · 2,896 · 2,815 · +81 · 3,129

Could only find part-time work · 1,267 · 1,263 · +4 · 1,200

Unemployed 27 weeks and over · 1,944 · 1,930 · +14 · 1,815

Marginally attached · 1,468 · 1,704 · -236 · 1,743

Discouraged workers · 414 · 441 · -27 · 555

Why people are unemployed, in thousands:

Reason · Sep 2026 · Aug 2026 · Change

Job losers and completed temporary jobs · 3,200 · 3,245 · -45

Job leavers · 741 · 914 · -173

Re-entrants · 2,289 · 2,137 · +152

New entrants · 818 · 702 · +116

Unemployment rates by group:

Group · Sep 2026 · Aug 2026 · Change

Total, 16 and over · 4.2 · 4.1 · +0.1

Adult men · 3.9 · 4.0 · -0.1

Adult women · 3.6 · 3.5 · +0.1

Teenagers · 14.5 · 14.1 · +0.4

White · 3.6 · 3.7 · -0.1

Black or African American · 7.0 · 6.0 · +1.0

Asian · 2.9 · 3.2 · -0.3

Hispanic or Latino · 4.7 · 4.8 · -0.1

Less than a high school diploma · 4.3 · 4.7 · -0.4

High school graduates, no college · 4.4 · 4.4 · 0.0

Some college or associate degree · 3.6 · 3.7 · -0.1

Bachelor's degree and higher · 2.5 · 2.7 · -0.2

Where The Jobs Came From

Health care and social assistance accounted for half of all private hiring:

Component · Change · Detail from the release

Ambulatory health care services · +13,400 · Offices of other health practitioners +6,700, dentists +4,000

Hospitals · +12,000 · Continuing its upward trend

Nursing and residential care facilities · -8,700 · Lost jobs over the month

Social assistance · +6,300 · Individual and family services +10,500

Health care and social assistance total · +23,000 · Health care alone +16,700, against a 12-month average of +33,000

Nonresidential specialty trade contractors · +12,300 · Residential specialty trades -7,900

Food services and drinking places · +10,800 · After an August beat that has since been revised lower

Merchant wholesalers, durable goods · +7,100 · n/a

Gasoline stations and fuel dealers · +7,300 · n/a

Plastics and rubber products · +4,600 · Manufacturing is up 72,000 since December 2025

Machinery manufacturing · +4,500 · n/a

Where jobs were lost:

Component · Change · Detail from the release

Employment services · -14,300 · Temporary help -10,900

Local government excluding education · -10,600 · Most of government's -17,000

Information · -10,000 · Publishing -4,000, broadcasting and content -3,000

Financial activities · -7,000 · Down 129,000 since May 2025, 90,000 of it in insurance carriers

Computer systems design and related services · -4,400 · Inside professional and business services

General merchandise retailers · -4,800 · Warehouse clubs and supercentres -4,100

Warehousing and storage · -4,100 · n/a

Accommodation · -3,600 · n/a

Information employment is 2,739,000, down 120,000 or 4.2% from a year ago. Computing infrastructure providers, data processing and web hosting are down 22,500 over the year to 453,800. Last month this brief described these job losses as the labour side of the AI capital-spending boom: firms buying capital and not hiring the staff that would once have come with it. September added 1,600 more.

How this month compares with ADP's report from Wednesday:

Sector · BLS, September · ADP, September

Total private · +46,000 · +90,000

Manufacturing · +9,000 · +17,000

Education and health · +20,000 · +55,000

Leisure and hospitality · +10,000 · +22,000

Financial activities · -7,000 · -16,000

Professional and business services · -9,000 · -11,000

ADP got the direction of every sector right and overstated four of the six. In August it understated private hiring by 89,000 on the first BLS estimate. This month it overstated it by 44,000.

Against This Morning's Open

• The Open defined the strong case as "headline well above 90K, unemployment at or below 4.1%, wages at 0.3% or above". All three came in on the other side: 29,000, 4.2% and 0.1%.
• It said a weak report "would confirm what is already mostly priced, so the reaction could be smaller," with October odds at about 26%. The print confirmed the weak case. Whether the market reaction was small is for the tape to show, and this brief does not quote a level.
• It also warned that this series "has missed by a wide margin recently". It missed again, by 61,000, and the revisions took back most of last month's upside surprise.

What This Sets Up

• Next Employment Situation: Friday 6 November, 08:30 ET, for October. It lands after the October FOMC meeting, so September is the last payrolls report the Committee sees before deciding.
• What would confirm the soft read: another three-month average near 50,000, a second private diffusion reading below 50, or further downward revisions to August and September.
• What would break it: a revision that restores the 81,000 private jobs removed from July and August, or wages back at 0.3% or more. Either would put the wage and employment argument for a hike back on the table.
• The 4.2% to watch: unrounded it is 4.18%, inside the 4.1%-4.3% range BLS says has held since March. A rise to 4.3% from more people losing jobs, not from more people entering the labour force, would be a different and more serious signal.

What Is This?

• What it is: The Bureau of Labor Statistics Employment Situation, released at 08:30 ET on the first Friday of most months. It combines two surveys. The establishment survey of businesses and government agencies produces payrolls, hours and earnings by industry. The household survey of about 60,000 households produces the unemployment rate, participation and the demographic detail. The two measure different things and often disagree.
• Why it matters: It is the most market-moving monthly release. Right now the Fed is in a hiking cycle (3.75%-4.00% after the 16 September hike) and focused on inflation, so the labour data matters for one question: is the labour market strong enough to justify another hike, or is it weakening enough to argue against one?
• How to read it: Read the revisions first, then the three-month average, then the headline. BLS says a monthly change of about 122,000 is statistically significant in the establishment survey and about 650,000 in the household survey, so most single months are within the noise. Each month's figure is revised twice in the following two releases. Wages are reported month on month and year on year; a rate near 0.3% m/m is roughly in line with a 3%-4% annual pace.

Want the full explainer? Economic News Events, Explained breaks down this release and every other one we cover, in plain English.

_For informational purposes only. Not investment advice._