Blog
Taft-Hartley Passed Over Truman's Veto and Still Shapes Every Drive
In June 1947 Congress overrode Harry Truman to rewrite labor law. The bans and permissions it added, from secondary boycotts to right-to-work, still decide organizing fights.
Harry Truman sent his veto of the Labor Management Relations Act to Congress on June 20, 1947, with a long message explaining, section by section, why he thought the bill was unworkable and unfair. The House did not bother to linger over his reasoning; it voted to override the same day. The Senate took a few more days and overrode on June 23. The new law took its popular name from its two sponsors, Senator Robert Taft of Ohio, the son of a president and the intellectual leader of congressional Republicans, and Representative Fred Hartley of New Jersey. Union leaders gave it other names, none of them printable in a convention program. Almost eight decades on, most Americans have never heard of the statute. Every union organizer in the country works inside the box it drew.
In short: The Taft-Hartley Act of 1947 amended the Wagner Act to restrict what unions could do: it banned the closed shop and secondary boycotts, let states pass right-to-work laws, removed supervisors from coverage and gave presidents power to halt strikes. Those rules, more than any single employer, explain why a modern organizing drive looks the way it does.
The year of the strikes
To understand why Congress was in such a mood, go back to 1946. With the war over and wartime wage controls lifting, prices jumped and workers who had held off on strikes during the war walked out in enormous numbers. Autoworkers struck General Motors. Steelworkers struck the steel industry. Coal miners under John L. Lewis struck, twice. Railroad workers threatened a national shutdown that so alarmed Truman, a president generally friendly to labor, that he went to Congress and asked for authority to draft striking railroad workers into the Army. That request was never enacted, but it shows how high tempers ran.
Public patience frayed. In the November 1946 midterm elections, Republicans won control of both houses of Congress for the first time since before the Depression. Many of the new members had campaigned against what they called union excesses, and conservative Southern Democrats were ready to join them. The Wagner Act of 1935 had been written almost entirely around employer misconduct; the new majority set out to balance it by writing in limits on unions.
It is worth saying plainly that some of the complaints had substance. A few unions did engage in jurisdictional strikes that shut down projects over which union's members would do a job. Some had locked out non-members through closed-shop agreements. Lewis had led wartime coal strikes that much of the public regarded as reckless. A law aimed at real abuses would have found support beyond the business lobby, and Taft-Hartley did. The question was whether the cure was proportionate to the disease.
What the act changed
The law runs to many pages, but a handful of provisions matter most to organizing today.
The closed shop was banned. Before 1947, a union contract could require an employer to hire only workers who were already union members. Taft-Hartley outlawed that arrangement. Unions could still negotiate a union shop, in which workers must join or pay fees after being hired, but the law later came to be read as requiring only payment of the costs of representation, not full membership.
States could ban union-security agreements entirely. Section 14(b) allowed any state to prohibit contracts that require workers to pay a union as a condition of employment. Those state laws are known as right-to-work laws. Several Southern and Plains states passed them quickly. Today 26 states have them, a number that fell by one when Michigan repealed its law in 2023.
Secondary boycotts were banned. A union in a dispute with one employer could no longer picket or strike a second, neutral employer to pressure the first. In practice, this rule broke one of labor's most effective tools: the ability of workers in one place to refuse to handle goods from a struck company, or to shut down a supplier until a customer settled.
Unions got their own list of unfair labor practices. Unions could no longer coerce workers in exercising their rights, refuse to bargain in good faith, or engage in certain featherbedding and jurisdictional strike tactics.
Supervisors and independent contractors were written out. Foremen had begun forming unions of their own. The law removed supervisors from its definition of employee, and it also excluded independent contractors.
Employers got a statutory free-speech clause. Section 8(c) said employer views and arguments are not unfair labor practices unless they contain threats of reprisal or promises of benefit. That clause is the legal footing for the anti-union meetings and literature that accompany modern campaigns.
Presidents got strike-stopping power. For strikes that threaten national health or safety, the president can seek an injunction imposing an eighty-day cooling-off period.
There was also a provision requiring union officers to sign affidavits swearing they were not Communists before their unions could use the board's processes. That requirement accelerated purges within the labor movement and was replaced in 1959.
How these rules play out on the ground
Consider how a drive at a large retail or logistics company works now. The employer is allowed, under Section 8(c), to tell workers repeatedly why it thinks a union would be bad for them, and for most of the law's history the board allowed mandatory meetings on paid time to deliver that message. Shift leads and team leads may be classified as supervisors and enlisted in the effort, because they are not covered employees and cannot be protected organizers. Workers at a contractor that runs the delivery vans may be outside the law entirely.
If the union wins and the employer stalls on a first contract, the most powerful pressure would be to draw in other workers: drivers who refuse to deliver, dockworkers who refuse to unload, workers at a supplier who stop work. Taft-Hartley makes most of that illegal. The union can strike its own employer, and can ask consumers to boycott, but solidarity across employers runs into the secondary-boycott ban almost immediately.
In a right-to-work state, the union that wins must still, by law, represent every worker in the bargaining unit fairly, members and non-members alike. Yet no worker can be required to pay for that representation. That gap, called the free-rider problem, is the core of the union argument against right-to-work. Supporters answer that no one should have to pay an organization as a condition of holding a job, and that a union that serves its members well will keep them. Both positions rest on a real value, individual freedom on one side and fairness in sharing costs on the other.
The cumulative effect is to localize labor disputes. The Wagner Act imagined workers acting together across workplaces. Taft-Hartley confined them, largely, to the employer whose name is on their paycheck, at a time when supply chains, franchising and subcontracting were about to make the name on the paycheck less and less meaningful.
A construction site with many employers
One fight over the secondary-boycott ban shows how a technical rule can reshape a whole industry. On a typical construction project, several contractors work side by side: a general contractor, an electrical firm, a plumbing firm, a drywall crew. Suppose one of those subcontractors is non-union, and a building-trades union pickets the entire site to protest. Is the picket aimed at the non-union firm, which the union has a dispute with, or at all the other contractors, which are neutral?
In 1951, in NLRB v. Denver Building and Construction Trades Council, the Supreme Court read the 1947 law to treat the general contractor and the subcontractors as separate employers, and found that picketing the whole site to force the non-union subcontractor off the job was an illegal secondary boycott. Construction unions have sought for decades to reverse that ruling through what they call common-situs picketing legislation. Congress actually passed such a bill in 1975, and President Gerald Ford, who had earlier signalled support for it, vetoed it in early 1976 after strong opposition from contractors and conservatives.
That episode is worth remembering because it complicates the usual partisan story. A Republican president seemed ready to sign a pro-union bill, then changed course under pressure. Unions have won and lost on labor law with presidents of both parties, and the details of a single rule, like whether a construction site counts as one workplace or many, can matter as much as any grand statement of principle.
Truman's veto, and what he did next
Truman's veto message argued the bill would increase strikes, intrude on the rights of employers and employees to bargain freely, and create endless litigation. He lost the override, but the veto became a centrepiece of his remarkable 1948 campaign, in which labor turned out heavily for him. Democrats regained Congress that year. Truman then asked for repeal, and the new Congress did not deliver it.
The irony is that Truman went on to use the law's emergency provisions himself several times, and presidents of both parties have done the same since. In 2002 George W. Bush invoked them to end a lockout at West Coast ports. A tool that unions denounced in 1947 became an ordinary feature of national labor relations.
Labor tried again in 1965, when President Lyndon Johnson and a House majority backed repeal of Section 14(b). The repeal passed the House but died in a Senate filibuster led by Everett Dirksen. That defeat set a pattern that has held for decades: labor-law changes that unions want can pass the House but stall in the Senate, while Section 14(b) remains.
The question the law never settled
Here is the argument this post wants to make. Taft-Hartley is often discussed as if it were a list of rules about unions. It is better understood as a decision about the scale at which workers may act. The labor movement's historic strength came from solidarity that crossed company lines: miners and railroad workers supporting each other, dockworkers refusing to handle struck goods, whole cities shutting down in general strikes, as several did in 1946. Congress decided that such breadth was a danger to the public and wrote rules to contain it.
There is a fair case for some of those rules. A neutral business with no say in a dispute can be harmed by a secondary boycott. Strikes that cut off coal or rail service can threaten the public. A law that only restrained employers would be lopsided in a different direction. But the scale problem is now acute. A warehouse worker at a third-party logistics firm in a right-to-work state, bargaining only with the subcontractor and unable to reach the retailer whose packages she handles, is in a much weaker position than the autoworker of 1946. The law's architecture fits an economy of large, vertically integrated employers. That economy is mostly gone.
That is why reform proposals keep returning to the same targets: overriding state right-to-work laws for private-sector workers, loosening the secondary-boycott ban, and narrowing the contractor and supervisor exclusions. Some reformers go further and propose sectoral bargaining, in which wages are set across an industry rather than one employer at a time, as some European countries do. Each of these would undo part of the 1947 bargain. Each would face the same Senate math that sank repeal in 1965. Follow the current proposals and the votes on them on our Congress page.
One more detail shows how durable the 1947 settlement has been. When Congress next amended federal labor law in a major way, with the Landrum-Griffin Act of 1959, it did not loosen the secondary-boycott ban. It tightened it, closing gaps that unions had found, while adding a bill of rights for union members and new reporting rules aimed at union corruption. The direction of travel, for most of the postwar era, ran one way.
Where the law could change next
The new Congress takes office on January 3, 2027, with Democrats in control of both chambers after the November 2026 elections. Whether that produces any change to federal labor law is an open question, not a forecast; the filibuster, the president's veto pen and the party's own divisions all stand in the way, as they have before. What is more certain is that most of the fights in 2027 will happen where they always have: in individual drives, at the National Labor Relations Board and in state legislatures, where right-to-work laws are passed and, occasionally, repealed.
If you want to see how often union-friendly measures have attracted votes from both parties, the record of bipartisan bills on the common ground page is a useful corrective to the idea that labor is a purely partisan cause.
What you can do
The rules of 1947 were written by Congress, and Congress can rewrite them. In the meantime, they can be worked around with care.
- Learn what your state allows. Whether your state is a right-to-work state changes how unions in it are funded and how they talk to workers. Our workers and unions topic page collects coverage of state fights over these laws.
- Track organizing where you live. The labor page lists the newest NLRB union elections and their results, so you can see which workplaces near you are organizing and whether drives in your industry are winning.
- Contact your senators about specific provisions. The secondary-boycott ban and Section 14(b) are concrete, nameable pieces of law. Find your senators through the representatives tool and ask where they stand on each, rather than on "labor law" in general.
- Back the workers who are bargaining under these limits. Local worker centres, strike funds and legal aid groups are listed by state on the organizations page.