Union Membership Statistics (2026)

10.0% of US wage and salary workers were union members in 2025, some 14.7 million people.

In 1983, the first year of comparable data, the rate was 20.1% and there were 17.7 million members. The rate has halved.

The member count fell by about three million over the same period while total employment grew substantially, so the decline in share is much steeper than the decline in bodies.

The public-sector union membership rate was 32.9% in 2025, more than five times the private-sector rate of 5.9%.

Local government is the most unionised part of the economy at 37.8%, which reflects the concentration of police officers, firefighters and teachers there.

Over the year the public-sector rate rose 0.7 percentage points while the private-sector rate was unchanged, so the gap widened slightly.

7.3 million union members work in the public sector and 7.4 million in the private sector, which is close to a fifty-fifty split.

That is the fact people find most surprising, and it follows directly from size. The private sector employs many times more people, so a 5.9% rate there produces roughly the same headcount as a 32.9% rate in government.

It also means that describing unions as a public-sector phenomenon is true about rates and misleading about people.

Utilities lead at 17.8%, followed by transportation and warehousing at 13.6% and educational services at 13.4%.

At the other end, finance is 0.8%, insurance 1.2%, professional and technical services 1.3%, agriculture 1.6% and food services and drinking places 1.8%.

The range from 0.8% to 17.8% within the private sector is wider than the gap between the private-sector average and the national average, which is why a single national figure tells you very little about any particular industry.

By occupation group, education, training and library occupations are highest at 32.5%, with protective service occupations close behind at 31.3%.

The lowest are farming, fishing and forestry at 2.4% and sales and related occupations at 2.5%.

Both of the top two are dominated by public employment, which is the occupational expression of the same public-private divide.

Men are at 10.4% and women at 9.6%. In 1983 those rates were 24.7% and 14.6%, so the gap has narrowed considerably, mostly because the male rate fell further.

Black workers remain the most unionised major group at 11.4%, ahead of White workers at 9.9%, Hispanic workers at 8.9% and Asian workers at 8.7%.

Age matters more than either. Workers aged 45 to 54 are at 12.6% while those aged 16 to 24 are at 4.7%, close to a threefold difference.

16.5 million workers, 11.2% of the workforce, were represented by a union in 2025.

That includes the 14.7 million members plus 1.8 million workers who report no union affiliation but whose jobs are covered by a union contract.

The distinction matters when reading any union statistic, because coverage and membership are different measures and the gap between them is over a million people.

Union members had median usual weekly earnings of $1,404 in 2025 against $1,174 for nonunion workers. Nonunion earnings were 84% of union earnings.

BLS states directly that these comparisons are on a broad level and do not control for many factors that can be important in explaining earnings differences, listing occupation, industry, age, firm size and geographic region.

That caveat is doing real work here. Unionised occupations skew toward public-sector professional roles such as teaching and protective services, which have different pay distributions for reasons unrelated to bargaining.

Full-time workers are unionised at 10.9% against 5.7% for part-time workers.

That roughly two-to-one gap compounds the earnings comparison above, because full-time work pays more per week by definition.

It is another reason the headline earnings gap should be read as a description of who is in unions rather than as a measure of what unions do to pay.

The 12.6% rate for workers aged 45 to 54 against 4.7% for those aged 16 to 24 is the sharpest demographic split in the release.

Part of that is tenure: union jobs in the public sector and utilities tend to have long service, and young workers are concentrated in retail and food service, two of the least unionised industries.

It also means the membership base skews older than the workforce, which has implications for the direction of the overall rate as those cohorts retire.

Across states, membership rates ranged from below 5% to nearly 25% in 2025.

BLS cautions that state union membership data are based on relatively few observations and only allow for broad characterisations.

For 2025 that caution is stronger than usual, for the reason in the next section.

The Current Population Survey for October 2025 was not collected because of the federal government shutdown.

As a result, 2025 annual estimates were produced using 11-month averages that exclude October, and BLS states they are not strictly comparable with annual averages for other years.

Union membership and earnings data come from only a quarter of the CPS sample, so they are already based on relatively few observations and are more affected than most series by a missing month. BLS specifically flags state estimates as needing caution.

A national rate of 10.0% is close to meaningless on its own, because the number that applies to any given worker ranges from 0.8% to 37.8% depending on sector and industry.

The earnings gap is real as a description and weak as evidence, because BLS says so in the release itself.

And the 2025 comparison to prior years should be made carefully, because a month is missing from the average.

Every figure is from the BLS news release Union Members -- 2025, which draws on the Current Population Survey.

The 1983 comparisons are BLS's own, from the first year for which comparable data exist.

The shutdown caveat is quoted from the box BLS placed at the top of the release rather than added by us.