Manufacturing Employment, Hours, and Earnings: 57 Data Points for 2025-2026

The Short Version

Manufacturing employment fell from 12.673 million in January 2025 to 12.611 million in July 2026. That is a net decrease of 62,000 jobs, or about 0.5%. The low point in this period was December 2025 at 12.580 million. Payrolls then added 31,000 jobs through July 2026, although the July reading is preliminary.

At the same time, the average manufacturing workweek rose from 40.0 hours to 40.4 hours. The increase looks small, but it equals roughly 20.8 additional scheduled hours per worker over a full year if sustained. A longer workweek alongside nearly flat payrolls can indicate that employers are meeting production needs with existing staff before making permanent hires.

Average hourly earnings increased from $34.73 to $36.87, a gain of $2.14 or 6.2%. Multiplying hourly earnings by weekly hours produces a simple gross weekly estimate of $1,489.55 in July 2026, compared with $1,389.20 in January 2025. That calculation is useful context, but it is not an official BLS weekly earnings series and does not include employer benefit costs.

Monthly Manufacturing Labor Data

The table uses seasonally adjusted Current Employment Statistics data. Seasonal adjustment removes recurring patterns such as holiday production, summer shutdowns, and model-year changeovers. It makes adjacent months easier to compare. Every employment value links to CES3000000001, every hours value to CES3000000002, and every earnings value to CES3000000003.

* July 2026 is preliminary. Source: BLS Current Employment Statistics, retrieved August 16, 2026 through the BLS public API.

What Happened to Manufacturing Employment?

Payrolls declined through the end of 2025

The employment series declined in every month from January through December 2025. The largest one-month drop in the displayed figures was 15,000 jobs, recorded from May to June and again from June to July. By December, manufacturing payrolls were 93,000 below January's level. Those are changes in the seasonally adjusted BLS payroll series, not a count of layoffs. Employment can fall through layoffs, retirements, quits that are not replaced, and establishment openings or closures.

Early 2026 brought stabilization, not a full recovery

Employment was nearly flat in January and February 2026, then rose to 12.598 million in March. After a small pullback, payrolls increased in June and July. The July level remained 62,000 below January 2025 and 14,000 below July 2025. A few positive months therefore do not erase the longer decline.

Why Hours Rose While Employment Fell

The workweek held near 40 hours through most of 2025, then reached 40.2 hours in January 2026 and 40.4 hours in April. It stayed at 40.4 through July. The BLS hours series covers average weekly hours paid to all employees in manufacturing. It is not a direct measurement of machine utilization, productive time, or overtime hours.

Manufacturers commonly adjust hours faster than headcount. Adding a Saturday shift or overtime can cover a short order surge without committing to another full-time employee. Cutting overtime can respond to a slowdown without layoffs. Because of that flexibility, weekly hours can serve as an early labor-demand signal, but one monthly move is not enough to call a trend.

The 0.4-hour increase between January 2025 and July 2026 equals 1% of the original 40-hour workweek. Applied mechanically to July's 12.611 million payroll jobs, it represents roughly five million additional paid hours in a week compared with a 40-hour schedule. That is an illustration, not an official BLS total-hours estimate, because averages and payroll counts have different survey and aggregation properties.

Plant managers should compare the national workweek with local overtime, backlog, absenteeism, open requisitions, and temporary labor. A 40.4-hour national average does not tell a specific shop whether staffing is adequate. It does offer a neutral outside reference for board reports, labor plans, and lender discussions.

Manufacturing Earnings Growth

Average hourly earnings rose in 17 of the 18 month-to-month comparisons shown. The only decline was from March to April 2025, when the average moved from $35.25 to $35.13. By December 2025, earnings reached $36.08. The July 2026 preliminary figure was $36.87, according to BLS series CES3000000003.

The 6.2% increase over the full displayed period is nominal. It has not been adjusted for consumer inflation, so it should not be read as a 6.2% gain in purchasing power. It is also an average, not the raise received by a typical worker. The average can change when the mix of jobs, industries, regions, experience levels, or overtime changes.

Average hourly earnings include regular pay and certain premium payments captured by the CES survey, but they are not total employer compensation. Health insurance, retirement contributions, payroll taxes, paid leave, and other benefits require a separate employer-cost source. For recruiting budgets, a shop should add those costs and local market premiums to the national wage reference.

For a quick scale check, July's $36.87 hourly average at 40.4 hours equals about $77,456 annualized over 52 weeks. That is not an official annual salary estimate. It assumes the weekly average continues all year and ignores unpaid time, bonuses, and changes in hours. Use it only to translate the hourly series into an intuitive order of magnitude.

How to Read the Three Measures Together

Employment shows the size of payrolls

CES employment counts people on nonfarm establishment payrolls who received pay for any part of the reference pay period. It excludes proprietors, the self-employed, unpaid family workers, private household workers, and agricultural workers. One person with jobs at two establishments may be counted twice. The BLS CES FAQ explains the scope.

Hours show paid labor intensity

Average weekly hours help distinguish a payroll change from a broader change in labor input. Falling employment with rising hours may mean the remaining workforce is covering more time. Rising employment with falling hours may reflect part-time hiring, less overtime, industry mix, or temporary scheduling changes. The series alone cannot identify the cause.

Earnings show price, with mix effects

Hourly earnings are a labor-price signal, but changes in who is employed also affect the average. If lower-paid jobs decline while higher-paid jobs hold steady, average earnings can rise without every worker receiving a raise. Pair earnings with occupational wage data and local offers before setting pay bands.

Together, the latest figures describe a market with fewer payroll jobs than in early 2025, a longer average workweek, and higher nominal hourly pay. That combination is consistent with cautious hiring and heavier use of existing labor capacity. It is not proof of a single cause. Orders, productivity, automation, retirements, reshoring projects, closures, and sector mix can all move the totals.

Uses for Manufacturers and Researchers

Workforce planning: compare your site's headcount and weekly hours with the national direction. If local hours are rising much faster while vacancies remain open, overtime fatigue and retention risk deserve attention. If payrolls grow while hours and output fall, examine onboarding, product mix, downtime, and demand.

Compensation reviews: use the hourly earnings trend as an outside reference, then add occupational, state, and metro data. National manufacturing averages blend production workers, engineers, managers, administrators, and many industries. They should inform a pay review, not dictate an offer.

Market research: machinery sellers, staffing firms, software vendors, and industrial service companies can use payroll and hours as demand context. More hours can increase maintenance, tooling, consumables, and safety needs even when employment is flat. A shrinking payroll can raise interest in automation and training, but vendor messaging should not claim causation from these data alone.

Economic reporting: always name the series, adjustment, unit, period, and preliminary status. “Manufacturing employment was 12.611 million in July 2026, seasonally adjusted and preliminary” is defensible. “Manufacturing created 12.6 million jobs” is wrong because the number is a level, not jobs created.

Methodology and Limitations

CES is a monthly establishment survey. BLS estimates employment, hours, and earnings using employer payroll records, then benchmarks employment estimates to administrative records. Details on sampling, estimation, nonresponse adjustment, seasonal adjustment, and annual benchmarking appear in the official technical notes.

The three series are seasonally adjusted. Seasonally adjusted values are best for month-to-month analysis. Researchers studying actual payroll levels in a particular month may need not-seasonally-adjusted series instead. Do not combine the two types in one change calculation.

Employment is in thousands. A displayed value of 12,611 means about 12,611,000 payroll jobs, not 12,611 people. Hours and earnings are averages for all employees. They do not describe the distribution, median worker, specific occupation, or individual establishment.

This page preserves the data vintage retrieved on August 16, 2026. BLS may revise recent months after publication and may revise a longer history during annual benchmarking. For the newest vintage, follow the linked series pages rather than treating this table as permanently final.

What This Data Does Not Tell You

National manufacturing statistics flatten substantial differences among industries. A semiconductor fab, meat processor, steel mill, cabinet shop, and aircraft plant operate with different capital needs, shift patterns, skill requirements, and pay structures. A stable national total can hide job growth in one manufacturing sector and contraction in another. Anyone making an investment or hiring decision should move from the national series to the matching three-digit or four-digit NAICS industry where available.

The figures also do not separate permanent employees from every contingent arrangement used inside a plant. Workers employed by a staffing agency are generally counted in the staffing industry's payroll data, even when their assignment is on a factory floor. A manufacturer can therefore add production capacity through temporary labor without the same increase appearing in manufacturing payroll employment. Outsourcing maintenance, logistics, inspection, or engineering work can create a similar industry-classification effect.

Geography matters too. National hourly earnings do not show what it takes to recruit a machinist in Wichita, a controls engineer in Detroit, or a maintenance technician near a Gulf Coast chemical cluster. Housing costs, commuting patterns, union coverage, competing employers, and the local supply of trained manufacturing workers affect market pay. Use state and metropolitan occupational employment and wage data for location-specific decisions.

Finally, these series do not measure output. A longer workweek may produce more goods, more work in process, or more downtime and rework. Employment can decline while production rises if productivity improves, or employment can rise faster than output during a ramp-up. Pair this labor dataset with Federal Reserve industrial production, Census shipments and orders, and BLS productivity measures before drawing conclusions about manufacturing industry performance.

Those limitations do not make the series weak. They define the job it can do. CES provides a timely monthly view of payrolls, paid hours, and employee earnings. It works best as a common baseline, followed by narrower operational and local evidence.

Building a Useful Factory Labor Dashboard

A shop can turn these national benchmarks into a better internal review by keeping its definitions consistent. Start with month-end headcount, average paid hours, overtime hours, open positions, quits, new hires, and base wage and salary expense. Add temporary labor separately. Then connect labor input to good units, standard hours earned, on-time shipments, scrap, rework, and unplanned downtime.

Compare year-over-year changes as well as month-to-month moves. Factory calendars create noise even after a local team accounts for holidays and shutdowns. A single large order, maintenance outage, or customer delay can distort one month. A three-month moving average can make the underlying direction easier to see without hiding a major break.

Keep wages and total manufacturing compensation separate. The BLS hourly earnings series is a wage measure. Your total employer cost also includes health coverage, payroll taxes, retirement contributions, paid leave, bonuses, and other benefits. A compensation survey or internal benefits report can fill that gap. Labeling the measures correctly prevents a wage benchmark from being mistaken for the loaded cost of labor.

Segment the dashboard where the business decision requires it. Production, maintenance, quality, engineering, warehouse, and office employees face different labor markets. Day, evening, night, and weekend shifts may carry different premiums. New hires and experienced workers may have different turnover patterns. One plant-wide average can conceal the exact group driving overtime or vacancies.

Use the BLS survey data as the outside comparison line, not a scorecard target. A 40.4-hour national workweek does not mean every plant should schedule 40.4 hours. The right local number depends on demand, capacity, safety, fatigue, training, and service requirements. The benchmark earns its place by prompting better questions.

Questions the Trend Raises for 2026

Will longer hours turn into hiring?

Four consecutive months at 40.4 weekly hours suggest manufacturers were using more paid time than they did in most of 2025. If orders remain firm, employers may eventually add shifts or headcount because sustained overtime carries fatigue, quality, safety, and retention costs. If demand softens, hours can fall quickly and hiring may never follow. Future monthly readings from BLS weekly hours will show which adjustment came first.

Has the payroll decline ended?

The employment level rose by 29,000 between February and July 2026, but June and July were preliminary when this report was updated. That increase is modest beside the 93,000-job decline during 2025. A convincing turn would require several months of gains that survive revision and appear across more than one manufacturing subsector. The annual benchmark can also alter the history.

Can earnings keep rising without cost pressure?

Higher earnings help plants recruit and retain skilled people, but labor cost per unit depends on output per hour as well as pay. A 6.2% increase in average hourly earnings does not translate automatically into a 6.2% increase in product cost. Productivity, overtime premiums, staffing mix, automation, yield, and the labor share of total cost determine the result. Manufacturers should connect compensation data to their own cost-per-good-unit measure.

What role will worker mix play?

Retirements and difficult-to-fill technical roles can change average pay even without uniform raises. Hiring more engineers, programmers, or experienced maintenance workers can lift the average. Losing lower-paid entry roles can do the same. Conversely, a rapid hiring wave in trainee positions can hold down average earnings while established workers still receive raises. Occupational detail is needed to separate wage growth from composition.

How should suppliers respond?

Industrial suppliers should treat the data as a signal for research, not an excuse for generic automation claims. A prospect with rising hours may care about uptime, quick-change tooling, preventive maintenance, contract labor, or training. Another may be deliberately running longer shifts because it is more efficient than adding capacity. Good outreach connects a verified plant problem to an economic result. It does not tell every factory that a national average proves what it needs.

Frequently Asked Questions

How many people work in US manufacturing?

The preliminary, seasonally adjusted estimate was 12.611 million payroll jobs in July 2026, according to BLS.

What is the average manufacturing workweek?

All manufacturing employees averaged 40.4 paid hours per week in July 2026, seasonally adjusted and preliminary, according to BLS.

What are average hourly earnings in manufacturing?

Average hourly earnings were $36.87 in July 2026, seasonally adjusted and preliminary, according to BLS. This average covers all employees and is not an occupational wage or total compensation figure.

Are these numbers revised?

Yes. Recent monthly estimates are revised as more reports arrive. Employment history is also benchmarked annually. Cite a retrieval date when reproducibility matters.

Sources

• BLS CES manufacturing employment series CES3000000001. Seasonally adjusted manufacturing payroll employment, in thousands.
• BLS CES manufacturing weekly hours series CES3000000002. Seasonally adjusted average weekly hours for all manufacturing employees.
• BLS CES manufacturing hourly earnings series CES3000000003. Seasonally adjusted average hourly earnings for all manufacturing employees.
• BLS Current Employment Statistics overview. Program releases, data tools, concepts, and notices.
• BLS CES frequently asked questions. Definitions, seasonal adjustment, revisions, and interpretation guidance.
• BLS CES technical notes. Survey design, estimation, benchmark revisions, and reliability.
• BLS public data API. Original machine-readable endpoint used to retrieve the three series.