Last updated: July 23, 2026

US Manufacturing Employment History: 2000-2026

American factories employed 12.582 million people in January 2026. That headline is useful, but the long view tells the real story. Manufacturing entered 2000 with 17.284 million jobs, lost nearly six million by 2010, rebuilt steadily through 2023, then softened again. This report puts every January reading in one source-linked table so journalists, analysts, economic developers, and manufacturers can cite the record without stitching together multiple downloads.

The complete 27-year employment table

The figures below are January seasonally adjusted payroll employment levels, measured in thousands. Total manufacturing comes from the BLS Current Employment Statistics series published through FRED. The two component columns use the corresponding durable goods and nondurable goods series. Durable and nondurable employment add to the total in every row.

What changed between 2000 and 2010

The steepest change in this record happened before the recovery began. Total factory employment fell from 17.284 million in January 2000 to 11.447 million in January 2010, a loss of 5.837 million jobs, or 33.8 percent. Durable goods producers accounted for most of that decline. Their payrolls dropped from 10.845 million to 6.976 million. Nondurable producers moved from 6.439 million to 4.471 million.

Two distinct contractions are visible. Payrolls fell rapidly from 2001 through 2004, then remained closer to 14 million for several years. The 2008 financial crisis triggered the second drop. Manufacturing entered January 2008 with 13.717 million jobs, stood at 12.552 million one year later, and reached 11.447 million in January 2010. Those annual snapshots do not show every monthly turn, but they clearly show the scale and speed of the contraction.

For anyone comparing modern hiring conditions with the early 2000s, this break matters. A plant can add workers while the national industry remains far below its earlier headcount. Employment is also not the same as output. Automation, process improvement, product mix, imports, exports, and hours worked can all move production without producing a one-for-one change in payroll jobs. Use this table to describe labor demand, not as a substitute for industrial production or manufacturing GDP.

The long factory jobs recovery

January 2010 marks the low point in this annual series. From there, total manufacturing employment increased in most years. It reached 12.265 million in 2015, 12.527 million in 2018, and 12.790 million in 2019. The gain from January 2010 through January 2019 was 1.343 million jobs, equal to 11.7 percent.

Durable goods drove more of the rebound. Durable payrolls rose from 6.976 million in 2010 to 8.025 million in 2019, an increase of 1.049 million. Nondurable manufacturing moved from 4.471 million to 4.765 million, adding 294,000. That difference is useful for sales planning. Suppliers serving machinery, metals, transportation equipment, electronics, wood products, and other long-lived goods faced a larger swing in available labor than vendors concentrated in food, chemicals, paper, apparel, and other nondurable categories.

The recovery did not restore the employment base to its 2000 level. By January 2019, factories still employed 4.494 million fewer people than in January 2000, based on the same BLS series. That gap is one reason raw job counts need context. The industry changed its footprint, technology, sourcing, and output per worker during the period.

The pandemic shock and 2023 peak

Annual January figures soften the visual effect of the short, severe 2020 shutdown because January 2020 preceded the broad disruption. The next January reading shows the damage: total manufacturing employment moved from 12.745 million in January 2020 to 12.142 million in January 2021. That is a loss of 603,000 jobs, or 4.7 percent.

Hiring returned quickly. Total employment reached 12.564 million in January 2022 and 12.903 million in January 2023. The 2023 figure was the highest January level in this dataset since 2008. Durable goods employment reached 8.036 million, while nondurable employment reached 4.867 million.

The post-pandemic peak still sat 4.381 million jobs below January 2000. Put another way, the industry recovered from its 2010 low and its 2021 disruption without returning to the headcount common at the start of the century. That is a sharper and more defensible statement than saying manufacturing employment simply rose or fell.

Why durable goods carry more employment

Durable goods represented 7.840 million jobs in June 2026, compared with 4.758 million nondurable jobs. In the January table, durable manufacturers consistently account for roughly three-fifths of factory employment. The category includes large labor markets such as fabricated metals, machinery, transportation equipment, computer and electronic products, and furniture.

Durable industries also tend to respond strongly to business investment, housing, vehicle demand, and interest rates. Orders can be postponed when customers become cautious. Nondurable demand often behaves differently because food, beverages, chemicals, petroleum products, paper, and similar goods are consumed or processed on shorter cycles. That does not make nondurable employment stable in every recession, but it helps explain why the two lines should be watched separately.

For a manufacturing marketer, the split is practical. A national total can hide opposing movements in customer segments. If durable payrolls weaken while nondurable payrolls hold steady, one generic campaign aimed at “manufacturers” may miss where budgets and hiring pressure actually sit. The BLS Current Employment Statistics program publishes detailed industry series that can support a narrower segment analysis.

How to read and cite this dataset

The unit is thousands of jobs

A table value of 12,582 means 12,582,000 payroll jobs. These are jobs reported by establishments, not a count of unique people. The BLS CES questions and answers explains that the survey measures nonfarm payroll employment, hours, and earnings using employer payroll records.

Every row uses January

Using one month makes the annual comparison consistent and keeps the table readable. It does not represent an annual average. If your claim concerns a full calendar year, download monthly observations from FRED's manufacturing employment page and calculate the 12-month mean. If your claim concerns the newest available month, use the latest observation shown on that page instead of the January snapshot.

The figures are seasonally adjusted

Seasonal adjustment removes recurring calendar patterns, making adjacent months easier to compare. The BLS seasonal adjustment documentation describes the process and annual updates. Data can be revised as more employer reports arrive and seasonal factors are recalculated.

Revisions are normal

CES estimates are benchmarked to unemployment insurance tax records. The BLS benchmark article explains why historical payroll estimates may change. This page records the values available on July 23, 2026. For publication, check the linked source and note your access date.

What the data means for manufacturers and suppliers

The 2000 to 2026 decline does not mean there are no hiring opportunities. It means the market employs fewer people overall and that gains have been uneven. A supplier selling recruitment, training, automation, maintenance, software, or production equipment should pair national employment with local establishment counts, job openings, wages, and sector production. Employment tells you how much payroll demand exists. It does not tell you which plant has an active project.

Recent softness also deserves precision. January manufacturing employment fell from 12.903 million in 2023 to 12.582 million in 2026, a reduction of 321,000 jobs. Durable goods accounted for 233,000 of that decline, based on the durable series. Nondurable goods accounted for 88,000, based on the nondurable series. The contraction is meaningful, but it is far smaller than either the 2001 to 2004 slide or the 2008 to 2010 loss.

For location strategy, do not use the national curve alone. A state can gain factories and workers while the country declines, or lose share during a national expansion. Use state and metro data from BLS Quarterly Census of Employment and Wages when a decision depends on a specific labor market.

Questions this employment history can answer

How many US manufacturing jobs have been lost since 2000?

The US had 4.702 million fewer manufacturing jobs in January 2026 than in January 2000. Employment decreased from 17.284 million to 12.582 million, a 27.2 percent reduction. Calling all 4.702 million “lost” can be misleading if it suggests that the same plants and occupations disappeared. The figure is a net change. New facilities opened, old facilities closed, companies changed industries, and employers added and eliminated positions throughout the period. The national total captures the combined result.

The starting month also matters. January 2000 was near the end of a long period in which manufacturing payrolls were much larger. Choosing the 2010 low instead produces a different story: employment was 1.135 million higher in January 2026 than in January 2010. Both comparisons are correct. A responsible citation states the dates and does not use one window to imply that the other trend did not happen.

When did US manufacturing employment bottom out?

Among the January readings shown here, the low was 11.447 million jobs in January 2010. Monthly data can identify a more precise turning point, and FRED lets readers display or download every monthly observation. This report uses January consistently because mixing monthly lows with annual snapshots would make the comparisons harder to reproduce.

The low followed two major periods of job reduction. The first stretched across the early 2000s. The second arrived with the financial crisis and recession. From January 2008 to January 2010 alone, manufacturing payrolls decreased by 2.270 million jobs. That two-year loss was larger than the entire net recovery from January 2010 through January 2026.

Did manufacturing employment recover after COVID-19?

Yes, measured against the January 2021 reading. Total employment rose from 12.142 million in January 2021 to 12.903 million in January 2023. The increase of 761,000 jobs more than restored the 603,000-job difference between January 2020 and January 2021. This annual comparison does not measure the deepest shutdown-month loss, so it should not be described as the full pandemic decline or recovery.

Employment then eased. The January total was 12.875 million in 2024, 12.673 million in 2025, and 12.582 million in 2026. June 2026 stood slightly higher at 12.598 million on the same source page. A claim about current conditions should use the newest monthly figure. A claim about comparable annual checkpoints can use this table.

What share of factory jobs are in durable goods?

Durable goods represented 62.0 percent of manufacturing employment in January 2026. That calculation divides 7.803 million durable goods jobs by 12.582 million total manufacturing jobs. Nondurable goods represented the remaining 38.0 percent, with 4.779 million jobs.

The mix has moved only modestly across the full period. Durable goods accounted for 62.7 percent in January 2000. Its slightly smaller 2026 share does not mean every durable industry followed the same path. Transportation equipment can behave differently from computers, primary metals, machinery, or furniture. Anyone sizing a particular supplier market should move from the two broad categories to detailed three-digit or four-digit NAICS data.

Does fewer employment mean the United States makes less?

Not necessarily. Employment is one input, while output measures the quantity or value of production. A factory can produce more with fewer workers when equipment, software, plant design, worker skills, product value, or scheduling improves. It can also add workers while output falls if hours decline, new staff are still training, or the product mix becomes more labor intensive.

For that reason, do not use the 27.2 percent employment decline as a manufacturing output decline. Pair labor data with the Federal Reserve industrial production and capacity utilization release or the Bureau of Economic Analysis GDP by industry accounts. Those systems answer different questions and use different units.

Are these workers, jobs, or survey responses?

They are estimates of payroll jobs. The CES program surveys businesses and government agencies, then estimates employment for the national economy and detailed industries. A person holding two payroll jobs can be counted twice. Owners who are not on a payroll, unpaid family workers, farm workers, and private household workers are outside the standard CES scope, as described in the BLS program questions and answers.

This distinction matters when comparing CES with the household survey. The household survey counts employed people and assigns each person to one job, while the establishment survey counts jobs. Differences between the two do not automatically indicate an error. They reflect different samples, concepts, and coverage.

Can this dataset be used in a news story or report?

Yes. Cite the original BLS series through the linked FRED pages and include the observation date, unit, seasonal adjustment, and access date. If you reproduce the full table, credit this page for the compilation and BLS for the underlying statistics. The suggested citation below gives editors a clean starting point.

Before publishing a high-stakes figure, open the relevant series page and confirm the value. Payroll statistics are revised. A later benchmark can produce a small difference from the value recorded here. That is expected behavior for an economic time series, not a flaw in the source.

Methodology

We downloaded the monthly MANEMP, DMANEMP, and NDMANEMP series and selected the January observation for each year from 2000 through 2026. Each series is measured in thousands of persons and seasonally adjusted. We calculated annual percentage changes from the displayed total employment values. No forecast values were added.

FRED republishes the series from the US Bureau of Labor Statistics and provides stable landing pages, charts, notes, and downloads. The source agency is BLS. The publisher used for each linked observation is the Federal Reserve Bank of St. Louis. Readers can reproduce the table using the CSV download option on each series page.

Our calculations retain the published precision of the employment series. Percentage changes are rounded to one decimal place only after comparing the underlying values. Totals may look rounded because the source unit is one thousand jobs. We did not interpolate missing months, combine different seasonal settings, or substitute household employment estimates. The durable and nondurable components were checked against the total for each displayed year. This approach keeps the report reproducible and avoids false precision. Researchers who need quarterly averages, recession troughs, or confidence intervals should work from the complete monthly download and document any additional calculation choices alongside the result.

Turn Your Manufacturing Website Into a Lead Engine

Get the free 5-day email course: the 5 fixes that turn your site from a brochure into an RFQ machine.

Get the Free 5-Day Course →

Cite this page

Sources

1. Federal Reserve Bank of St. Louis, Manufacturing Employment (MANEMP), source: US Bureau of Labor Statistics.
2. Federal Reserve Bank of St. Louis, Durable Goods Manufacturing Employment (DMANEMP), source: US Bureau of Labor Statistics.
3. Federal Reserve Bank of St. Louis, Nondurable Goods Manufacturing Employment (NDMANEMP), source: US Bureau of Labor Statistics.
4. US Bureau of Labor Statistics, Current Employment Statistics overview.
5. US Bureau of Labor Statistics, CES questions and answers.
6. US Bureau of Labor Statistics, CES seasonal adjustment.
7. US Bureau of Labor Statistics, CES benchmark article.
8. US Bureau of Labor Statistics, Quarterly Census of Employment and Wages.
9. Federal Reserve Board, Industrial Production and Capacity Utilization.
10. US Bureau of Economic Analysis, GDP by Industry.