This essay appears in print in Work, Inequality, Basic Income.

In December 2005 more than 30,000 New York City transit workers walked out over economic issues despite the state of New York’s Taylor Law, which prohibits all public sector strikes. Not only did the workers face the loss of two days’ pay for each day on strike, but a court ordered that the union be fined $1 million per day. Union president Roger Toussaint held firm, likening the strikers to Rosa Parks. “There is a higher calling than the law,” he declared. “That is justice and equality.”

The transit strike exemplified labor civil disobedience at its most effective. The workers were not staging a symbolic event; they brought the city’s transit system to a halt. They claimed their fundamental right to collective action despite a statute that outlawed it. For a precious moment, public attention was riveted on the drama of workers defying a draconian strike ban.

How did national labor leaders react?

Union decline matters. For half a century, it has moved in lock step with the increase in income inequality.

AFL-CIO president John Sweeney issued a routine statement of support, while most others did nothing at all. To anybody watching the drama unfold, the message was clear: there is no right to strike, even in the House of Labor.

About a decade earlier in 1996, Stephen Lerner, fresh from a successful campaign to organize Los Angeles janitors, had warned in Boston Review that private sector unions faced an existential crisis: density could soon drop from 10.3 percent to 5 percent if unions did not expand their activity beyond the limits imposed by American law. He called for unions to develop broad organizing strategies—industry-wide and regional—and to engage in civil disobedience. Few embraced these radical strategies. Today private sector union density is about 6.5 percent, not quite as low as Lerner predicted, but down from a high of over 30 percent in the mid-1950s.

Union decline matters. For half a century, it has moved in lock step with the increase in income inequality. According to an International Monetary Fund study of twenty advanced economies, union decline accounted for about half of the increase in net income inequality from 1980 to 2012. In the heyday of American unionism, CEOs made about 25 times the annual compensation of the average worker; today, the multiple is more than 350. Meanwhile, as Thomas Edsell and others have warned for decades, the decline of unions has deprived the Democratic Party of its strongest link to white workers. The overwhelming majority of unions continue to endorse Democratic candidates (including Hillary Clinton in the 2016 election), but with ever-diminishing effect.

Until two decades ago it was possible to blame union decline on backward labor leaders, such as George Meany, who were so steeped in business unionism that they could not see the need to organize broadly, much less to ally with other social movements across lines of race, gender, and immigration status. Since then, however, we have seen continued shrinkage under leaders who are, for the most part, well intentioned and savvy.

Organized labor is being strangled by laws that block workers from exercising the rights to organize, to strike, and to act in solidarity.

The problem is structural. National union officials are not well positioned to lead a challenge to corporate power. Institutions with big treasuries and tit-for-tat relations with establishment politicians cannot be expected to undertake risky and polarizing actions. Although leaders might see the need to build working-class power, the immediate incentives all point toward the narrow needs of their particular union’s members. This constraint is rooted in the American system of exclusive representation, which divides workers into thousands of bargaining unit boxes, gives unions property interests in particular boxes, and penalizes unions for doing anything other than defending existing boxes and acquiring new ones.

The prospects for union revival may seem bleaker than ever during the Trump administration, even as the triumph of right-wing populism makes more urgent what was already apparent: the need to build a labor movement that can fight for the interests of the working class in the face of corporate power.

But prospects are not as grim as they appear. Over the past decade, there has been an undeniable shift toward class politics, most visibly evidenced by Occupy Wall Street, the Bernie Sanders campaign, the Fight for Fifteen, and the rise of a Black Lives Matter movement that supports economic justice demands, including the right to organize. Building the labor movement in this period of danger and opportunity will require not only heeding Lerner’s call for a strategic shift and extralegal action; labor must also reclaim the right to strike and confront the deep structural disabilities that impede unions from challenging corporate power.

As Lerner diagnosed twenty years ago, U.S. labor law blocks unions and workers from effective organizing and striking. Then as now, the law’s protections for workers’ rights amount to little more than paper guarantees, while its restrictions are downright deadly. Indeed the Committee on Freedom of Association of the International Labor Organization (ILO) has held that the United States is violating international standards by failing to protect the right to organize, by banning secondary strikes and boycotts across the board, and by allowing employers to permanently replace workers who strike. The ban on secondary strikes is especially debilitating, because it prevents workers who have economic power, such as organized grocery workers, from aiding workers who do not, for example unorganized packing house workers. If the grocery workers support striking packers by refusing to handle food packed by strikebreakers, they are said to be engaging in an illegal secondary strike.

The prospects for union revival seem bleaker than ever, even as Trump makes more urgent the need for a labor movement that can fight for the working class in the face of corporate power.

But the law cuts even deeper, deforming workers’ organizations at their inception. As amended by the Taft-Hartley Act of 1947 (tagged by unionists as the “Slave Labor Law”), the National Labor Relations Act (NLRA) confronts workers with a choice between two inadequate forms of organization: statutory “labor organizations,” popularly known as unions, and “others,” for example workers’ centers that organize outside the statutory framework. At first glance, the choice seems obvious. Only unions can demand and engage in collective bargaining. But unions are subject to so many restrictions that some workers’ organizations (such as the Restaurant Opportunities Centers United) are willing to forego collective bargaining in order to avoid them, while others (including the Coalition of Immokalee Workers) consider themselves lucky to be excluded from the NLRA altogether. In the 1960s Cesar Chavez of the United Farm Workers rejected NLRA coverage for farm workers on the ground that it would inscribe “a glowing epitaph on our tombstone.”

The obvious response would be to reform the law. But labor faces a double bind: American workers have never won a significant piece of workers’ rights legislation without first engaging in exactly the kind of strikes and other forms of noncooperation that current labor laws forbid. The Erdman Act of 1898, the Clayton Act of 1914, the Railway Labor Act of 1926, the Norris-LaGuardia Anti-Injunction Act of 1932, the Wagner Act (NLRA) of 1935, and the public sector collective bargaining laws of the 1970s were all preceded by dramatic strikes and mass disobedience.

By comparison, organized labor’s more recent legislative campaigns all failed despite Democratic ascendancy in both houses of Congress and the White House. The Labor Law Reform bill of 1978, the striker replacement bills of the early 1990s, and the Employee Free Choice Act (EFCA) of 2007–9 succumbed to a combination of tepid presidential support (Carter, Clinton, and Obama to labor leaders: “I’m with you; just wait until I’ve spent my political capital on other things”) and the filibuster. Even if enacted, those bills would have provided only modest protections for workers’ rights, well short of the far-reaching changes necessary to reverse union decline. Given the booming influence of money on politics, the skewed representation in the Senate, and the gerrymandered House, we simply cannot expect ordinary politics to produce the reforms that would give unions a fighting chance of revival. Organizing, it seems, must precede legislation.

The Service Employees International Union (SEIU) is the only big union to launch the kind of confrontational campaign urged by Lerner. For the past four years, SEIU has poured money and organizers into the nationwide Fight for Fifteen campaign. With its combination of sectoral organizing and civil disobedience, Fight for Fifteen has scored a number of victories, including the enactment of fifteen-dollar minimum wage laws in several jurisdictions as well as the inclusion of a fifteen-dollar minimum wage plank in the Democratic Party platform. The campaign has gained SEIU few dues-paying union members—which to some critics earns it a failing grade—but it has validated organized labor as a champion of low-wage workers and accelerated the shift toward class politics.

It should come as no surprise that Fight for Fifteen has made more progress on wages than on union growth. Employers have always resisted unionization far more tenaciously than wage increases. They understand that unionism entails a workplace regime shift, while wage increases merely redistribute wealth for a time. Conversely, organized labor has never achieved major growth without prioritizing the rights to organize and strike above economic gain. The Fight for Fifteen and—for that matter—most of the labor movement’s activity, would be far more effective if it were tied to a long-term strategy for winning three core rights for workers: rights to organize, strike, and act in solidarity. Lacking those rights (whether de facto or officially), the movement will be of little use in struggles for social justice or in alliances with other movements. The labor movement of the early twentieth century, which propelled unionism to its historic high, grasped this point. Even the cigar-chomping business unionists of Samuel Gompers’s era seized on opportunities to trumpet the constitutional rights to organize and strike, sometimes in support of open lawbreaking by leftist unions and workers.

American workers have never won a significant battle without first engaging in exactly the kind of strikes that current labor laws forbid. Organizing always precedes legislation.

In order to win workers’ rights, organized labor should act like a rights movement. History tells us that rights movements—from abolition to women’s suffrage to civil rights—succeed when they claim a few key rights, exercise them at every opportunity, and place them front and center in every phase of movement activity, including organizing, protest, civil disobedience, legislative advocacy, administrative advocacy, and litigation. Not only does this kind of focus help to sway public opinion, but also—perhaps more importantly—it assures adherents and convinces opponents that the movement is serious. No workers contemplating extralegal exercise of labor rights should doubt that the movement will come to their support and that they are participating in a historic struggle for rights that will be carried through to victory.

How can workers claim their rights in defiance of duly enacted laws? Social movements typically answer this kind of question with reference to higher law, especially the Constitution. For example, the civil rights movement defied Jim Crow in the name of the Constitution’s equal protection clause. The labor movement of the early twentieth century held that anti-strike laws established “involuntary servitude” in violation of the Thirteenth Amendment, while anti-picketing and anti-boycott laws transgressed the First Amendment freedoms of free speech and association. Neither movement waited for courts to recognize their rights; they interpreted the Constitution for themselves.

International norms also protect the rights to organize, strike, and act in solidarity. A tremendous advance would be to bring U.S. labor law into compliance. In the meantime workers are fully justified in deploying tactics of peaceful disobedience in the course of organizing, striking, and acting in solidarity.

What would it mean in practice for labor to act like a rights movement? It would not mean that unions ride back to glory on the slogan of workers’ rights. Far more likely, struggles would continue to center on substantive demands, such as a fifteen-dollar minimum wage or a union contract. But a long-term commitment to workers’ rights would entail basic changes not only in tactics, but conceivably in the very definition of unions as government-anointed exclusive representatives.

Peaceful disobedience and political action would be two key components of a rights-centered strategy. When people think of civil disobedience today, most think of symbolic protests or brief disruptions designed to attract public attention. Unions have conducted some important actions of this type, for example during the San Francisco hotel strike of 2010 and the more recent Fight for Fifteen.

Employers resist unionization more tenaciously than wage increases. They understand that unionism entails a workplace regime shift, while wage increases merely redistribute wealth for a time.

This kind of action is a big advance, but it does not go to the heart of the matter. There is a reason why the workers’ movement, alone among progressive social movements, has been able to sustain mass organizations over a long period of time: the capacity of workers to withhold labor. Symbolic actions and consumer boycotts might extract concessions from employers who are vulnerable to public pressure—for example, fast food and retail enterprises—but even they are unlikely to accept unions (as opposed to wage hikes) solely on the basis of consumer pressure. By contrast the sit-down strikes of the 1930s succeeded because they shut down production, not only in mass production industries such as automotive and rubber manufacturing, but also in service industries such as retail and laundry.

We propose that everyone who is concerned about union revival make it a top priority to support and publicize civil disobedient exercises of workers’ rights wherever and whenever they happen, especially where workers are violating restrictions on the rights to organize, strike, or act in solidarity. For example, in response to the spontaneous worker occupation of the Republic Windows factory in 2008, the United Electrical Workers Union (UE) promptly moved to support the workers and maximize the impact of their courageous action instead of worrying about possible union liability or negative reactions from politicians or employers.

A rights movement can also gain ground by campaigning for rights legislation, even if its bills fall short of passage. Consider the twentieth-century labor movement, which won all of our major national workers’ rights statutes. Strikes were no more popular then than they are now. Yet for decades unions campaigned for the total abolition of anti-strike, anti-organizing, and anti-boycott laws and injunctions. Many bills were defeated, but each gave unionists an opportunity to demand the rights to organize and strike under the First Amendment freedom of association and the Thirteenth Amendment ban on involuntary servitude. And, although the provisions of the bills varied, the focus on rights remained strong until Congress finally passed the Norris-LaGuardia Anti-Injunction Act of 1932 and the National Labor Relations Act of 1935, which protected the rights to organize and strike until eroded by judicial decisions and the Taft-Hartley Act of 1947.

Compare that to the campaign for the Employee Free Choice Act of 2008–9 (EFCA), organized labor’s only major effort to reform labor law in the past two decades. Right down to the title, every feature was shaped to nest comfortably in the prevailing labor law regime and dominant (anti-labor) politics. The bill declared the stirring principle that workers should have a “Free Choice” between unbridled employer domination and the crabbed version of unionism decreed by Taft-Hartley. And on the right to strike, the bill added another constraint: compulsory arbitration of first contracts. This might have made it easier for unions to sign up dues-paying members in the short run, but it would have impeded the long-run struggle. With compulsory arbitration of first contracts, organized labor would have had a hard time convincing anyone that, as Richard Trumka once put it, unions must have “their only true weapon—the right to strike.”

If EFCA is an example of what not to do, then what should organized labor be doing? Endorsed by the NAACP as well as the AFL-CIO, the Employee Empowerment Act of 2014 (EEA) would be a solid step forward. It authorizes workers who have suffered anti-union discrimination to bring a direct federal court action against their employers. In other words, it treats anti-union discrimination as a civil right, making available the remedies of the Civil Rights Act of 1964, including injunctions, compensatory damages, and punitive damages. As Tom Geoghegan, Richard Kahlenberg, and Moshe Marvit observe, the bill offers a concrete opportunity to raise and defend the proposition that labor subordination is of equivalent moral importance to racial and gender subordination. At the state level, unions might work to create and strengthen state wage boards, which, as Kate Andrias has pointed out, offer opportunities to develop sectoral bargaining outside the Taft-Hartley framework.

So far we have talked about rights, but not about the legally structured institutions that shape the exercise of those rights. Those institutions are fatally flawed, and organized labor needs to confront that reality now, before it loses even more ground defending them.

The unionism decreed by the NLRA should be rebuilt from scratch. Although we tend to think today of the 30-plus percent union density of the 1950s as the good old days, it was not. If it becomes politically possible to amend the NLRA to make possible union revival, then it will also be possible to jettison the act’s crabbed definition of unions themselves.

In order to win workers’ rights, organized labor should act like a rights movement. Peaceful disobedience and political action would be two key components of a rights-centered strategy.

At the heart of the difficulty lies the system of exclusive representation. Unions that enjoy the government-conferred status of exclusive representative have little incentive and few legal avenues to build the movement as a whole.

As amended by the Taft-Hartley Act, the NLRA carves workers up into government-defined “bargaining unit” boxes, anoints a single union as the exclusive representative of the workers in any particular box, and restricts that union to bargaining over “terms and conditions of employment,” a category that does not include a host of issues of vital concern to workers, including plant closings, automation, and control of pension funds. Unions stand on relatively solid legal ground when they attend to the immediate self-interest of workers in a single box, but risk employer retaliation and legal sanctions if they act on the view that the fortunes of all workers rise and fall together. For example, the flat ban on secondary boycotts blocks the workers in each bargaining unit box from acting in solidarity with workers in other boxes.

But it gets worse. Unlike corporations, which must compete in the marketplace to retain their investors, Taft-Hartley unions enjoy government-conferred monopolies over their workers. The fact that union busters repeat this point ad nauseum does not make it any less true. Once a union establishes itself as the exclusive representative in a bargaining unit, it extinguishes the freedom of workers in that unit to shift their allegiance to another union except through an arduous process of “decertification” that presents the employer with a golden opportunity to dispense with unions altogether. Union democracy can provide workers with considerable control in some settings (especially single-facility local unions, sites of some of the most vigorous popular democracy anywhere in the United States), but the law gives national union leaders enormous latitude to suppress or avoid democracy.

This kind of power presents union leaders with an almost irresistible temptation to offer, in the words of journalist and union veteran Bob Fitch, “solidarity for sale” to employers and politicians. When a union achieves the status of exclusive representative, it takes ownership of the workers’ right to strike. From that point on, the union may trade the right away, and—even if it does not—the workers may be fired for striking without the union’s approval. (Compare France, where the right to strike belongs to workers, not unions, and is often exercised in support of class-wide demands.) Even the most militant labor leaders typically accept a blanket no-strike clause in exchange for stability in their bargaining unit boxes, whence all dues flow. As a result most union workers are prohibited by contract from striking during all but the window periods between contracts. Because contracts expire at various times, sympathy strikes and political strikes are effectively precluded.

Exclusive representation encourages the sale of solidarity not only to employers, but also to politicians. For a vivid illustration, we need look no further than the response of labor leaders to Bernie Sanders’s presidential campaign. Although neutral during the campaign, Trumka acknowledged after the fact that Sanders “elevated critical issues and strengthened the foundation of our movement.” The overwhelming majority of national labor leaders, however, bowed down to the candidate backed by Wall Street, Hillary Clinton. Given the system of exclusive representation, it made sense to curry favor with the likely winner.

Exclusive representation opens the door to special restrictions on labor rights as well. Unionists routinely complain that unions are denied constitutional rights enjoyed by other voluntary associations, for example the right to engage in secondary picketing and political boycotts. But exclusive representation gives courts a plausible response, namely that because government confers the special privilege of exclusive representation on unions and not other associations, it can impose special restrictions as well.

Exclusive representation causes unions to have little incentive and few legal avenues to build the labor movement, and encourages the sale of solidarity to employers and politicians.

Finally, exclusive representation undermines organized labor’s claim that unionism serves as a vehicle not only for higher wages, but also for industrial democracy. At the time of Lerner’s article two decades ago, nobody imagined that anti-labor interest groups would launch a successful cultural offensive against unionism, presenting themselves as defenders of democracy in the fight over EFCA (Save our secret ballot!), and of workers’ constitutional rights in the campaign for the “right to work” without paying union dues.

For most unionists, resistance to the “right to work” is almost as instinctive as respect for picket lines. If a workplace is unionized, employees must pay union dues. This is justified as necessary to solve the free-rider problem. But there are plenty of solutions to that problem that do not involve forcing workers to pay dues to a union that owns their bargaining unit box solely because it mustered majority support at some point in the past. What if, for example, workers had to pay dues but could decide which union should receive the money? Similar systems have been implemented in France and Italy, where employers bargain over wages at the national level with the most representative union in the industry, but other unions compete with that union and can displace it if workers so choose.

In a political context where unions are facing an existential threat, questioning exclusive representation might seem academic. But it is not. In Only One Thing Can Save Us (2014), Tom Geoghegan suggests that the movement might consider bargaining away exclusive representation in exchange for rights protections such as those offered by the Employee Empowerment Act. One might disagree about the particulars, but this is the kind of discussion that the movement needs now, not only to shape the long-term campaign for legal reform, but also to inform organizing today.

If, as expected, the Supreme Court under a Trump administration strikes down the union shop in the public sector, organized labor will be presented with a challenge and opportunity to develop new systems. In states where labor remains strong, experimenting with alternatives might be possible—for example, requiring workers to pay a representation fee, but giving them a choice of organizations.

Everyone who is concerned about union revival must support and publicize civil disobedient exercises of workers’ rights wherever and whenever they happen.

In the private sector we might put more energy into members-only or non-majority unionism. As Charles Morris has shown, a strong legal case can be made that the NLRA requires employers to bargain with non-majority unions over the wages and conditions of their members only. Earlier this year NLRB Member Kent Hirozawa agreed. And for years observers have predicted that the Board would reverse its current rule that, in the absence of an exclusive representative, a worker can be fired for insisting on bringing a representative into a disciplinary meeting. But with or without assistance from the Board, non-majority unionism offers promising opportunities to build unions outside the system of exclusive representation. After all, this was the standard path to union recognition prior to NLRA. As the best organizers testify today, their work is inevitably a process of building power in the face of employer resistance.

We have presented a simple argument. Organized labor is being strangled by laws that block workers from exercising the rights to organize, to strike, and to act in solidarity. Unions should respond by building a rights movement, placing the struggle for those rights front and center in all movement activity, including organizing, protest, civil disobedience, political action, administrative advocacy, and litigation. In the process, the system of exclusive representation must be challenged, and labor must develop an alternative that permits broad solidarity and promotes worker freedom. We offer this proposal as a contribution to what we hope will be a productive discussion about how best to move forward in this moment of crisis and opportunity.