Union Membership Statistics 2026: 10.0% of US Workers, 71% Approval, and a $230-a-Week Pay Gap

10.0%
US union membership rate, 2025 (BLS)
14.7M
union members, up 410,000 on 2024 (BLS)
7.1
Americans who approve of unions for every one who belongs to one (our metric, section 7)
$11,960
raw annual median pay gap, union vs non-union full-time workers (our calculation)

Union membership is one of the oldest continuous labour-market series in the United States, and one of the most misquoted. Headlines mix up members with workers covered by a contract, compare years the Bureau of Labor Statistics itself says are not comparable, and quote the "union wage premium" without saying whether it is a raw gap or a controlled estimate. This page puts the official numbers in one place: the BLS annual union release for 2025, the BLS major work stoppages series, Gallup's long-running approval question, NLRB election data, and OECD and UK figures for comparison.

Every figure was checked against the primary publication on 16 November 2026, and each one is labelled with its source, period and population. Where we computed something ourselves โ€” the pay gap by occupation, the approval-to-membership ratio, the share of members in the public sector โ€” it is marked as our calculation and the arithmetic is shown.

Union membership cheat sheet

The benchmarks most often needed for an article, a board paper or a labour-relations brief. Full sources are in the methodology section.

BenchmarkFigureSource and scope
Union membership rate10.0% (2024: 9.9%)BLS, wage and salary workers, 2025
Union members14.7 million (14,665,000)BLS, 2025
Membership rate in 198320.1% (17.7 million)BLS, first year of comparable data
Represented by a union16.5 million (11.2%)BLS, members plus covered non-members, 2025
Public sector rate32.9% (7.3 million)BLS, 2025
Private sector rate5.9% (7.4 million)BLS, 2025
Most unionised industry (private)Utilities, 17.8%BLS, 2025
Most unionised occupation groupEducation, training and library, 32.5%BLS, 2025
Median weekly pay, members vs non-union$1,404 vs $1,174BLS, full-time workers, 2025
Highest / lowest stateHawaii 24.8% / South Dakota 2.3%BLS, 2025
Public approval of unions71%Gallup, August 3โ€“24, 2026, 1,200 US adults
Want unions to have more influence47% (record)Gallup, 2026
Major work stoppages begun30, idling 306,800 workersBLS, 2025, 1,000+ workers
NLRB union elections held1,498 (โˆ’30%), win rate 69.8%CAP analysis of NLRB data, calendar 2025
OECD average union density15% (1985: 30%)OECD/AIAS ICTWSS, 2023/24
UK union density22.4% (6.6 million)UK DBT, employees, 2025

1. How many workers are in a union

In 2025, 14,665,000 US wage and salary workers were union members, 10.0% of the 146.7 million employed, according to the BLS "Union Members โ€” 2025" release of February 18, 2026. That was 410,000 more members than in 2024 (14,255,000, 9.9%). BLS describes the rate as "little changed".

A further 1.8 million workers were not members but were covered by a union contract, so 16,486,000 workers โ€” 11.2% โ€” were represented by a union. Represented is the better number when the question is "how many people have their pay set by collective bargaining"; members is the right number for "how many people belong to a union".

2. The long-run decline

BLS's comparable series starts in 1983, when 20.1% of workers (17.7 million) were union members. By 2025 the rate had halved to 10.0% and the number of members had fallen by about 3 million, even though employment grew by tens of millions. Averaged over the 42 years, the rate lost about 2.4 percentage points a decade (our calculation).

US union membership rate, 1983 vs 2024 vs 2025

Percent of employed wage and salary workers, BLS. Bars drawn to scale on 0โ€“25%.

1983
20.1%
2024
9.9%
2025 (11-month average)
10.0%

Membership by sector tells the story of the decline: it happened almost entirely in private industry. Private-sector members (7.4 million) and public-sector members (7.3 million) are now roughly equal in number, even though the private sector employs more than five times as many people.

3. Public vs private sector, and by industry

In 2025 32.9% of public-sector workers were union members, against 5.9% in the private sector โ€” 5.6 times higher. Within government, local government had the highest rate (37.8%), followed by state (29.6%) and federal (27.2%). The public-sector rate rose 0.7 points on the year, while the private-sector rate was unchanged.

Union membership rate by industry, 2025

Percent of employed wage and salary workers, BLS Table 3. Bars drawn to scale on 0โ€“40%.

Local government
37.8%
State government
29.6%
Federal government
27.2%
Utilities
17.8%
Transportation and warehousing
13.6%
Educational services (private)
13.4%
Construction
11.1%
Motion pictures and sound recording
11.0%
Manufacturing
7.7%
Health care and social assistance
6.8%
Retail trade
4.0%
Accommodation and food services
2.4%
Finance
0.8%

By raw numbers the biggest private-sector union industries are health care and social assistance (1.37 million members), manufacturing (1.11 million) and construction (995,000). Transportation and warehousing fell from 15.8% in 2024 to 13.6% in 2025, the largest drop among major industries; construction rose from 10.3% to 11.1%. BLS notes that 2025 industries use the 2022 Census classification, another reason to treat year-on-year moves with care.

By occupation group, the highest rates were in education, training and library occupations (32.5%) and protective service occupations (31.3%); the lowest in farming, fishing and forestry (2.4%) and sales and related occupations (2.5%).

4. Who belongs: age, sex, race and hours

Group2024 rate2025 rate2025 members
Men10.2%10.4%7.9 million
Women9.5%9.6%6.8 million
Ages 16โ€“244.3%4.7%900,000
Ages 25โ€“348.8%8.9%3.05 million
Ages 35โ€“4410.8%11.0%3.66 million
Ages 45โ€“5412.6%12.6%3.57 million
Ages 55โ€“6411.8%12.0%2.75 million
Ages 65+9.1%8.4%732,000
Full-time10.7%10.9%13.2 million
Part-time5.7%5.7%1.4 million

By race and ethnicity, Black workers had the highest membership rate in 2025 (11.4%), followed by White (9.9%), Hispanic (8.9%) and Asian workers (8.7%). The youngest workers are the least unionised: at 4.7%, people aged 16โ€“24 are less than half as likely to be members as those aged 45โ€“54.

5. Union vs non-union pay

Among full-time wage and salary workers in 2025, union members had median weekly earnings of $1,404, against $1,174 for workers not represented by a union. BLS phrases it the other way round: non-union workers earned 84% of what union members earned. Workers represented by a union (members plus covered non-members) had a median of $1,386.

The pay gap by occupation (our calculation)

Comparing members and non-union workers within the same occupation group strips out some of that mix. The gap is large in skilled trades and protective services, small in professional jobs, and negative in three groups.

Occupation group (full-time, 2025)Union medianNon-union medianGap
Protective service$1,728$1,011+70.9%
Construction and extraction$1,561$1,004+55.5%
Installation, maintenance and repair$1,571$1,140+37.8%
Transportation and material moving$1,174$866+35.6%
Education, training and library$1,483$1,130+31.2%
Production$1,161$945+22.9%
Office and administrative support$1,145$963+18.9%
Healthcare support$884$777+13.8%
Food preparation and serving$757$699+8.3%
Healthcare practitioners and technical$1,605$1,521+5.5%
Management$1,862$1,850+0.6%
Sales and related$1,043$1,116โˆ’6.5%
Computer and mathematical$1,665$2,045โˆ’18.6%

Medians from BLS Table 4 (2025); the gap is union รท non-union โˆ’ 1. BLS publishes no union median for farming, fishing and forestry.

Unions and pay gaps between groups (our calculation)

Among non-union workers, women's median weekly pay was 81.9% of men's ($1,056 vs $1,289); among union members it was 87.2% ($1,297 vs $1,487). Hispanic workers' median was 76.8% of White workers' outside unions and 84.9% inside them. The pattern does not hold for every group: Black workers' median relative to White workers was 80.6% outside unions and 79.5% inside, and Asian union members earned less than Asian non-union workers ($1,445 vs $1,589). For more on pay gaps, see our pay transparency statistics.

The "union premium" is not one number. Among full-time workers it runs from +71% in protective services to โˆ’19% in computer and maths jobs โ€” the occupation mix matters more than the headline.

6. Union membership by state

State rates ranged from 24.8% in Hawaii to 2.3% in South Dakota in 2025 โ€” a 10.8-fold difference. Four states were at 18% or above: Hawaii, New York (21.3%), Alaska (18.1%) and Washington (18.0%), with Connecticut and Rhode Island next at 16.1%. Four were below 3%: South Dakota, North Carolina (2.5%), South Carolina (2.7%) and Arkansas (2.8%).

Highest rates, 2025RateLowest rates, 2025Rate
Hawaii24.8%South Dakota2.3%
New York21.3%North Carolina2.5%
Alaska18.1%South Carolina2.7%
Washington18.0%Arkansas2.8%
Connecticut16.1%Utah3.8%
Rhode Island16.1%Arizona4.1%
Oregon15.1%Idaho4.1%
California14.9%Mississippi4.2%

By number of members, California (2.49 million) and New York (1.83 million) lead by a wide margin, followed by Illinois (758,000), Texas (673,000) and Washington (632,000).

7. Public opinion: approval at a 60-year high

Gallup has asked "Do you approve or disapprove of labor unions?" since 1936. In its August 3โ€“24, 2026 survey of 1,200 US adults (margin of error ยฑ4 points), 71% approved and 21% disapproved. That ties 2022 and matches 1965; the last higher readings were 73% in August 1959 and 74โ€“75% in 1953 and 1957. The low point was 48% in August 2009.

Approval of labor unions, selected years

Percent of US adults who approve, Gallup. Bars drawn to scale on 0โ€“80%.

1957 (January)
75%
1979
55%
1999 (August)
65%
2009
48%
2016
56%
2022
71%
2025
68%
2026
71%
  • By party: 89% of Democrats, 70% of independents and 52% of Republicans approve.
  • By age: 77% of 18โ€“34-year-olds approve, against 67% of those 55 and older.
  • More influence: 47% want unions to have more influence than today, 26% the same and 23% less โ€” the highest "more" since the question began in 1999. Among 18โ€“34s it is 52%; among Republicans, 19%.
  • The future: 39% expect unions to get weaker, 29% to stay the same and 27% to get stronger (34% said stronger in 2023).
  • Union households: 17% of adults say they or someone in their household is a union member; 7% say they personally are, plus 2% where both they and another household member belong.

8. Strikes and work stoppages

BLS counts "major" work stoppages โ€” strikes and lockouts involving 1,000 or more workers and lasting at least one full shift. 30 major work stoppages began in 2025, idling 306,800 workers, according to the release of February 20, 2026. The 2006โ€“2025 average is 17.8 a year; the series high was 470 in 1952 and the low 5 in 2009.

  • Government led: 17 stoppages were in state and local government (six local, 11 state) and 13 in private industry. Government stoppages involved 232,800 workers, private ones 74,000.
  • Education and health dominated: 196,500 workers, 64.0% of those idled, were in education and health services. Manufacturing accounted for 6,200 (2.0%).
  • Days idle: 711,600 in government stoppages and 784,400 in private industry, about 1.5 million in total (our sum).

9. Union elections and organising

Organising activity cooled in 2025 after several years of growth. A Center for American Progress analysis of NLRB certified-election data (February 11, 2026) found 1,498 NLRB-overseen union elections in 2025, down 30% from 2,124 in 2024, with about 59,000 fewer workers voting (โˆ’42% from 142,000). Unions won 69.8% of elections. On a fiscal-year basis, employees and unions filed about 2,100 RC (representation) petitions in FY 2025, the fewest since FY 2022, according to an Ogletree Deakins summary of the NLRB's annual statistics.

Elections are the narrow end of the funnel: a single election can cover a few workers or thousands, and the BLS membership count also moves with employment in already-unionised workplaces. That is how membership could rise by 410,000 in 2025 while elections fell.

10. The US in international context

The OECD's September 2025 report on the OECD/AIAS ICTWSS database finds that union density across OECD countries has halved since 1985, from 30% to 15% in 2023/24. Rates range from 4.7% in Colombia, 5.6% in Estonia and 7.4% in Hungary to more than 60% in Sweden and Denmark and 90% in Iceland. Unionisation is far higher in the public sector (41.3%) than the private (10.1%), and nearly equal for women (14.2%) and men (14.9%).

Collective bargaining coverage โ€” the share of employees whose terms are set by an agreement โ€” fell from 47% in 1985 to about a third (the report gives both 33.6% and 33.5% for 2023/24). The comparable US figure, workers represented by a union, is 11.2%.

Country / groupUnion densitySource and year
IcelandAbout 90%OECD, 2023/24
Sweden, DenmarkMore than 60%OECD, 2023/24
United Kingdom22.4% (public 48.5%, private 12.1%)UK DBT, employees, 2025
OECD average15%OECD, 2023/24
United States10.0% (public 32.9%, private 5.9%)BLS, 2025
Colombia (lowest)4.7%OECD, 2023/24

In the UK, the Department for Business and Trade's bulletin of May 28, 2026 reports that 22.4% of employees were union members in 2025, up from a record-low 22.0% in 2024, with membership up 192,000 to 6.6 million. Density was 32.4% in 1995. Women (25.2%) are more likely to be members than men (19.6%), the reverse of the US; education is the most unionised industry (47.4%); Northern Ireland has the highest density (32.4%) and Inner London the lowest (17.9%). DBT estimates the union wage premium at 5.3%.

11. What the numbers mean for employers

Know whether your rewards are bargaining subjects

Under the NLRA, an employer with a recognised union may not change wages, hours or other mandatory subjects of bargaining without bargaining first. Recognition schemes that work like pay โ€” scheduled bonuses, cash-equivalent gift cards โ€” are worth reviewing with labour counsel before launch or change.

Expect a mixed workforce

One in nine covered workers is not a member, and many employers run union and non-union sites side by side. Reward rules should be explicit about who is eligible and why, so a programme does not look like a selective inducement.

Benchmark pay by occupation, not by headline

The raw union gap ranges from +71% to โˆ’19% by occupation. If you compete for trades, drivers or protective-service workers, the union rate is a real market reference; in tech and sales it is not.

Listen before an election does it for you

With 71% public approval and 77% among under-35s, the conditions that drive organising โ€” pay, scheduling, voice โ€” deserve attention on their own merits. Stay interviews and visible recognition are cheaper than turnover.

For the wider picture, see our layoff statistics, salary increase statistics, frontline and deskless workforce statistics and retention strategies ranked by evidence.

Key takeaways

  • Membership is stable at a low level: 10.0% in 2025, half the 1983 rate, with 14.7 million members.
  • Unions are now a public-sector institution: 15% of workers, 50% of members.
  • The pay gap is real but uneven: $230 a week overall, +55% in construction, negative in tech.
  • Geography is extreme: Hawaii is 10.8 times as unionised as South Dakota; two states hold 29% of members.
  • Opinion runs far ahead of membership: 71% approve, 47% want more union influence, 10% belong.
  • Strikes are public-sector led: 17 of 2025's 30 major stoppages, and 76% of workers idled.

Recognition that works for every worker

Rewordin lets HR teams send digital gift cards in 150+ countries and local currencies, with eligibility rules set by group, position or level โ€” so union and non-union sites can follow the policy you agree on.

Methodology and sources

Every figure on this page was checked against the primary publication on 16 November 2026. BLS union data come from the Current Population Survey and cover employed wage and salary workers aged 16 and over; the 2025 figures are 11-month averages excluding October 2025. Earnings are usual median weekly earnings of full-time wage and salary workers. Gallup figures are from the final topline of its August 2026 Work and Education survey, including its crosstabs by party and age. NLRB election figures are from the Center for American Progress's analysis of NLRB certified elections and an Ogletree Deakins summary of the NLRB's FY 2025 petition statistics; we could not open the NLRB's own petition table and say so. We did not use vendor or aggregator "union statistics" pages whose sources we could not check.

  • US Bureau of Labor Statistics, "Union Members โ€” 2025", news release USDL-26-0229, February 18, 2026, and Tables 1 (by age, sex, hours), 2 (median weekly earnings), 3 (by industry and sector), 4 (earnings by occupation) and 5 (by state).
  • US Bureau of Labor Statistics, "Major Work Stoppages in 2025", news release, February 20, 2026.
  • Gallup, Gallup Poll Social Series: Work and Education, final topline and crosstabs, August 3โ€“24, 2026 (questions 12โ€“14 and D19).
  • Center for American Progress, Aurelia Glass, "NLRB-Overseen Union Elections Fell in 2025 Amid Trump Administration Attacks", February 11, 2026.
  • Ogletree Deakins (Jesse R. Dill, Zachary V. Zagger), "Three Takeaways From the NLRB's FY 2025 Petition Statistics", National Law Review, January 30, 2026.
  • OECD, "Membership of Unions and Employers' Organisations, and Bargaining Coverage: Standing, but Losing Ground", September 2025 (OECD/AIAS ICTWSS database).
  • UK Department for Business and Trade, "Trade Union Membership, UK 1995 to 2025: Statistical Bulletin", May 28, 2026.
  • National Labor Relations Board, "Bargaining in Good Faith with Employees' Union Representative (Section 8(d) & 8(a)(5))".
  • Derived metrics (ours, recomputable from the tables above): the raw union pay gap overall, by demographic group and by occupation (union median รท non-union median โˆ’ 1; annual = weekly ร— 52); covered non-members (represented minus members); the public sector's share of employment (public members รท public rate) and of members; the California + New York share; the state ratio; the approval-to-membership ratio; the average size and days idle per worker of major stoppages; and the decline per decade since 1983. BLS inputs are rounded, so derived ratios carry small rounding error.

    About the authors

    MK
    Maciej Kamieniak
    Founder & CEO, Rewordin

    Maciej is the founder and CEO of Rewordin, a global employee rewards and recognition platform delivering gift cards in 150+ countries. He works directly with HR, People Ops and finance teams on reward and retention programmes and writes about the research behind them. Based in Wrocล‚aw, Poland. Connect on LinkedIn โ†’

    NK
    Natalia Kamieniak
    CFO, Rewordin

    Natalia is the CFO of Rewordin and co-reviewer of every quantitative claim published on the platform โ€” including the occupation pay-gap table in section 5 and the approval gap in section 7, both published with their inputs rather than as survey findings. Connect on LinkedIn โ†’

    Last reviewed: 16 November 2026 ยท Date published: 16 November 2026 ยท Next data update: BLS union release (expected early 2027), Gallup Labor Day poll (August 2027)
    Every statistic on this page is attributed to a named source with its reference period and population. Derived metrics are labelled and their arithmetic shown.

    What percentage of US workers are in a union?

    10.0% of wage and salary workers were union members in 2025, 14.7 million people (BLS, February 18, 2026). Including non-members covered by a union contract, 11.2% โ€” 16.5 million โ€” were represented by a union.

    Is union membership increasing?

    Slightly in number, not meaningfully in rate. Members rose by 410,000 to 14.7 million in 2025 and the rate moved from 9.9% to 10.0%, which BLS calls little changed. The long-run trend is down: the rate was 20.1% in 1983.

    Do union workers earn more?

    On a raw comparison, yes: full-time union members had median weekly earnings of $1,404 in 2025 against $1,174 for non-union workers, a 19.6% gap. It varies widely by occupation, from +71% in protective service to โˆ’19% in computer and mathematical jobs, and it is not a controlled estimate.

    Which state has the highest union membership?

    Hawaii, at 24.8% in 2025, followed by New York (21.3%), Alaska (18.1%) and Washington (18.0%). South Dakota is lowest at 2.3%. California has the most members, 2.49 million.

    Which industries are most unionised?

    Government above all: local government 37.8%, state 29.6%, federal 27.2%. In the private sector, utilities (17.8%), transportation and warehousing (13.6%) and educational services (13.4%) lead; finance (0.8%) is lowest.

    What percentage of Americans support unions?

    71% approved of labor unions in Gallup's August 2026 survey, tying 2022 as the highest since 1965. 47% want unions to have more influence, a record for that question.

    How many strikes were there in 2025?

    BLS counted 30 major work stoppages (1,000+ workers) beginning in 2025, involving 306,800 workers; 17 were in state and local government. Smaller strikes are not included in the BLS count.