The Bureau of Labor Statistics reported that 10.0 percent of wage and salary workers were union members in 2025, little changed from the prior year. That is 14.7 million people. In 1983, the first year with comparable data, the rate was 20.1 percent and there were 17.7 million members, from a considerably smaller workforce.
Underneath that single figure sits a split that explains most of the argument. Public-sector membership was 32.9 percent in 2025, up seven-tenths of a point over the year. Private-sector membership was 5.9 percent, unchanged. Public employees are more than five times as likely to belong to a union as private employees, and that ratio has widened for decades.
The gap that drives the policy debate is between what people tell pollsters about unions in general and how many actually belong to one. Approval of labor unions has polled near its highest levels since the 1960s, while membership has continued its slow decline. Both things are true at once, and each side reads the gap differently.
Supporters of easier organizing rules treat it as evidence that procedural barriers, not worker preference, explain the decline: if people approve of unions but do not have one, something between the preference and the outcome is doing the work. Opponents argue that approving of unions in the abstract is a different question from wanting one at your own workplace, with your own employer and your own pay at stake, and that surveys measure the first far more reliably than the second.
Unionization varies enormously by sector. In 2025 the highest rates were in utilities at 17.8 percent, transportation and warehousing at 13.6 percent, and educational services. The lowest were in financial activities at 1.5 percent, professional and business services at 2.1 percent, and leisure and hospitality.
That distribution matters for the policy argument because it maps onto which parts of the economy are growing. The heavily unionized sectors are largely mature or shrinking as a share of employment, while the fastest-growing service and professional categories start from almost no union presence. Supporters of rule changes argue this is why organizing law matters more now than when the law was written; opponents argue it reflects what workers in those sectors want.
Most of the legislative argument concentrates on how a union gets recognized. Under current federal practice, a union that gathers enough signed authorization cards can petition for a secret-ballot election run by the National Labor Relations Board. Between the petition and the vote, employers may hold mandatory meetings with employees to present their case.
Card check would let a union be recognized once a majority signs cards, without a subsequent ballot. Several states have moved against it: a Mississippi law effective July 2026 bars employers seeking economic development incentives from recognizing a union based solely on signed cards where a Board-run election is available, and from disclosing employee contact information to a union without written consent.
The left generally argues that the election window gives employers a structural advantage, that remedies for illegal firings are too weak to deter them, and that the private-sector figure of 5.9 percent reflects the legal environment rather than worker demand. The right generally argues that a secret ballot protects workers from pressure by organizers as well as employers, that card check removes that protection, and that the public-private gap reflects competitive discipline rather than legal obstruction.
One point both sides tend to accept: the National Labor Relations Act was written in 1935 for continuous employment at a single worksite, and it maps awkwardly onto an economy with contract work, multiple employers and remote arrangements. They disagree entirely about what should replace it.
Want the core arguments from both sides, side by side?
See both sides of the Union Organizing Rules debate →