The Trajectories of Trade Unionism: Comparative-Historical Context

To understand how labor regulation in the United States came to assume the form of law, we need to start at a more abstract level of analysis — at the level of capitalism itself. Furthermore, it is possible to understand the particularities of the US case only by situating it within its broader historical and comparative context. To that end, this section starts by setting trade union formation and organization within its capitalist environment. It then traces the trajectories of trade union organization — surely an important dimension of a broader process of class formation — across different countries. Finally, I show that differences in trade union organization across countries are linked with different regimes of labor law. I identify three: an Anglophone model of dependency, a Scandinavian model of self-regulation, and a continental European model of displacement.

Capitalism and Competition

Capitalist competition creates contradictory tendencies for the process of working-class formation. As Marx and Engels remarked in The German Ideology, “Competition separates individuals from one another, not only the bourgeois but still more the workers, in spite of the fact that it brings them together.”¹ At the center of these challenges are the significant differences between capitalist firms. Capitalist firms differ substantially, both between and within industries, in terms of their organization of work, technology, and capital investment. These differences lead to substantial differences in productivity, output, and the ratio of capital to labor, which in turn create the possibility for significant wage differentials between workers of otherwise identical skill, experience, or training.² How unions respond to these facts, including the ways those responses are shaped by their social, economic, and institutional history, says much about how successfully unions will be able to transcend trade union sectionalism.³

The old, craft unions of the American Federation of Labor (AFL), for example, achieved gains for a narrow section of privileged workers in part by excluding other workers from job opportunities. Exclusion was achieved through “job control”: union apprenticeships ensured that the craft union controlled the knowledge required to perform the job, job territories protected that skill within the workplace by limiting the employer’s control over technology and work organization, and the closed shop limited access to jobs by mandating that employers only hired those workers trained by the union. By contrast, unions in Scandinavia and other parts of Europe took a solidaristic approach to labor organizing. This meant the rejection of a property-like claim over the introduction of new technology or work organization, and instead a reappropriation of higher productivity for all workers through the principles of “equal pay for equal work” and “taking wages out of competition.”

Solidaristic unions formed on a consistently industrial basis — meaning not just organizing workers of all skill levels, but also organizing industries rather than firms, labormarkets rather than workplaces. Thus, it is often mistakenly assumed that the emergence of the Congress of Industrial Organizations (CIO) in 1935 represented a genuine break with its craft-union predecessors. In truth, the CIO unions simply adapted different elements of the structures and strategies of the AFL craft model. As functional equivalents to skill and apprenticeship, the CIO unions substituted seniority rights and promotion ladders; in place of the closed shop, they substituted the union shop; and rather than relying on the external labor market, which a skill monopoly made possible, they relied on internal labor markets.⁴ Thus, while the CIO unions did seek to organize workers regardless of skill, they remained prisoners of the workplace-centered model of their forbearers, never bargained with employers on a consistently industrial level, and, however much they were able to challenge the tyranny of the workplace, never mounted a serious challenge to the tyranny of the market.⁵

The next section will make this analysis even more concrete, and look at both the origins and implications of various degrees of trade union sectionalism across countries and time. Chief among origins will be (1) the timing of capitalist development and (2) the union’s coordinating capacity — roughly, power, but more precisely the allocation of internal union authority over matters such as finances and strike decisions. On this point, I note that rates of union membership, or union density, while certainly an important source of union power, are not sufficient to grasp its full nature.⁶ In terms of implications, I will show how differences in these two dimensions led to different forms of regulating labor relations.

Dependency

Not surprisingly, trade unions first appeared where capitalism also developed earliest — in Britain. British colonial rule exported this nascent mode of production and its associated trade union forms across the globe, so that trade union organization and behavior in Britain is more similar to that of Australia, New Zealand, and even the United States than it is to those countries (France, Belgium, and the Netherlands) just across the Channel. “In the British Empire and to a lesser extent in the USA, they imposed their pre-existing attitudes and modes of behaviour on the new social environment.”⁷

Early capitalist development had several implications for trade union organization. Because capitalist relations of production preceded large-scale factory production, capitalist control over the labor process was limited, continuity with the earlier guild traditions remained strong, and markets were oriented locally.⁸ Consequently, the nature of work was skilled and these craftsmen enjoyed a high degree of control over the labor process. These features generated a form of trade union organization that was centered on the workplace (or, the “point of production”) and decentralized, with local union organizations retaining a high level of autonomy with respect to finances and workers’ collective action. This form of organization has provided the template for all subsequent Anglophone trade union organization, including its “industrial union” phases.

This workplace-centered, decentralized model of trade union organization had important implications — above all for the regulation of labor relations, in addition to much else. The first important implication is that this model produced an enormously adversarial set of workplace relations. Decentralized union strategies, premised on the assertion of job control, strike at the center of the capitalist’s control of the labor process. This produced an intense battle between unions and employers for the personal loyalties of the worker.⁹ Unfortunately, being rooted “at the point of production” did not necessarily produce more inclusive forms of organization.¹⁰ This was because, circumscribed as it was by firm-level bargaining organization, it left workers subordinate both to the firm and, consequently, the imperatives of the market. Firm-level bargaining is incapable of superseding the imperatives of the market in any meaningful way. As such, workers became in fact more dependent on employers, more closely identifying their interests with the success of the firm rather than with wider sections of the working class.

Second, and in many ways simply the flip side of the first, firm-level bargaining and trade union organization came at the expense of greater coordinating capacity at the industry and national levels. We have already seen the different forms of this in the previous section. Here, it is worth pointing out why a simple metric like union density, important as it is, is nevertheless insufficient for understanding union power. At the firm level, decentralized bargaining could in fact produce very high levels of density — close to 100 percent in firms that were organized, even in “right to work” states. Unfortunately, union density was zero in unorganized firms. By contrast, where unions organize and bargain at industrial and national levels, such as in many European countries, all firms in a certain sense are organized and there is a smaller difference between firm-level and national-level density. Such maldistribution of power and resources explains much about labor’s vulnerability and inability to respond effectively to a myriad of strategic and tactical problems. For instance, with large differences between organized and unorganized firms and sectors, issues like technological change become an existential threat to unions and their negotiated gains. In countries where unions bargain on an industry or national level, by contrast, technological change becomes an opportunity, through increased productivity, for wage increases and further working-class gains.

These two characteristic features of Anglophone trade unionism — a high level of industrial conflict and a weak coordinating capacity — led to what I will call a “dependency model” of labor-relations regulation. Where unions were unable to regulate labor relations themselves because they lacked the coordinating capacity to do so, and yet where they still created immense industrial conflict, the state stepped in to regulate labor relations. Unfortunately, under this tutelage union progress became dependent on the state. And as a relation of dependency, unions also became all the more vulnerable to shifts in economics and politics. In the next section, we will trace the development of this model as it took shape in the US, as a form of procedural dependency. In the remainder of this subsection, I will situate the US between two different versions of the Anglophone dependency model: Australia, where unions were in important ways even more dependent on the state, and Britain, where they were in important ways less.

Beginning in 1904 and lasting until roughly 2003, Australian labor relations were governed by a stunningly comprehensive system of compulsory conciliation and arbitration. Like its US and British counterparts, “the Australian union movement was rooted in the craft tradition of union organization”¹¹ and industrial conflict was an important impetus for the system of regulation the country adopted. This system forced employers to recognize unions and abide by the “awards” delivered by tribunals to settle disputes. These awards were not collective bargaining agreements negotiated between unions and employers, like we see in the US; rather, they were the decisions of government tribunals, arrived at through a quasi-judicial process. These awards specified hundreds of details of the employment contract, and covered union and nonunion workers.

From a certain perspective, this system looks like an enormous triumph for labor and the results, for workers, were impressive: comprehensive union representation of the workforce, and a substantial drop in wage dispersion. But unions had to give up a lot to secure it. It affected their form (the system was “federal,” in the same way that that means “national” in the US), the content of their demands (the tribunals limited the subjects of bargaining to those specified in law), and their internal governance. Most fundamentally, they had to accept that there was no right to strike. It also left the Australian labor movement extraordinarily dependent on the state. The political right attacked this system in 1996 and then entirely dismantled it in 2004. By 2014, union density had declined by more than half, from 31 percent in 1996 to 15 percent in 2014, a more precipitous decline than any other labor movement in the advanced capitalist world, save New Zealand.¹² And the decline in New Zealand was for identical reasons: the political dismantling of its own, closely related compulsory arbitration system. Since then, Australia, along with other Anglophone countries, such as Canada and New Zealand, has converged on the US’s Wagner model of dependency.

At the opposite end of the dependency model is Britain. Owing to a high level of union density and working-class consciousness, British unions developed a legal system called collective laissez faire, where unions enjoyed broad freedom of action, and the state and law, both procedurally and substantively, were kept at bay. This system came quite close to replicating the model of self-regulation developed in Scandinavia. Given its craft-union history, however, they were unable to forge the state-independent institutions required for self-regulation along with employers.¹³ Furthermore, in areas where unions had weak influence, government-established trade boards — very similar to Australia’s model of compulsory arbitration — set industry-specific wages and working standards.¹⁴ In addition, since the Thatcher reforms of the 1980s, the British model has converged closer to the US model of procedural dependency.

Self-Regulation

For an entirely different approach to the regulation of labor relations, we can look to Scandinavia. Unlike the Anglophone countries, capitalism developed later in countries such as Denmark and Sweden.¹⁵ Because capitalist development came later, unions did not develop the same decentralized, workplace-based mode of organization as in Anglophone countries. In this case, the appearance of the factory coincided with the emergence of capitalist relations of production, making the job-control unionism of the Anglophone type an anachronism. Even in Denmark, where unions still bear the imprint of craft organization, the exception proves the rule: bargaining and collective action is highly coordinated and more centralized.

These initial conditions led to a form of union organization with high coordination capacity and union-based workplace representation. In this later-industrializing environment, inclusive strategies of solidarity rather than exclusive strategies of job control were essential. Large-scale factory production erased the differences between skilled and unskilled workers, making job-control strategies of exclusion useless. Therefore, union power, if were to exist, had to come from solidarity rather than segmentation. Unions were thus compelled early on to take more coordinated action, typically striking successive employers in “whipsaw” actions. Employers responded by organizing coordinated, multiemployer lockouts. Out of these conflicts at the beginning of the twentieth century, voluntary basic agreements between the two peak organizations of the employers and unions established a framework for comprehensive self-regulation. A central feature of these agreements is a “peace obligation,” an agreement not to strike, during the term of the agreement. Crucially, the framework agreements make the peak organizations responsible for ensuring that collective bargaining agreements are observed and carried out by “all affiliated organizations.” This may appear as a constraint on labor’s freedom of action but, as we will see, it also prevents the defection of privileged sectors of the labor movement that would undermine the “taking wages out of competition” that centralized bargaining can achieve. Note also that centralization does not necessarily imply a lack of union democracy, as rank-and-file workers in Denmark “have ample opportunities to influence the choices of union leaders, including the right to ratify or reject collective bargaining agreements through binding referenda.”¹⁶ Finally, although the locus of union authority is centralized at national and industry level, above the firm, workplaces are well organized and workers represented by union organizations — an important distinction from the works-council forms of representation in France and Germany, considered next.

It is difficult to understate the surprisingly limited role of the state in this regime. For instance, the setting of wages, hours, and employment terms are the exclusive domain of labor and capital — the state has no role. To underscore this, not only is there no statutory minimum wage in Sweden or Denmark, but labor unions have historically opposed it unequivocally. Unlike the US and other Anglophone countries, the use of economic weapons is largely unregulated, and although strike rates are low by international standards, both employers and unions make use of secondary and sympathetic actions when necessary. Also unlike Anglophone countries, no procedure exists for the government to “recognize” a union, and the state’s role in supervising the collective bargaining process is otherwise limited or nonexistent. When anti-union corporations Toys ‘R Us and McDonald’s set up shop in Sweden, for example, union recognition was achieved neither by a government-supervised representation election (as in the US) nor by compulsory state-based arbitration (as in Australia). Rather, unions used tactics prohibited in the US, such as secondary actions, to blockade these employers, who eventually had no option but to relent. Furthermore, as I have indicated, the main framework governing the relationship between labor and capital in both Denmark and Sweden is not a statute, but a “basic agreement” — a private but encompassing “contract,” if you will — between the main federations of employers and employees.

Finally, the state does play a role in regulating Scandinavian labor relations, but primarily as a means of legal backing for these agreements. Thus, in both countries labor courts exist to adjudicate disputes between employers and unions. Although clothed with “public” power and authority, which makes their decisions binding on organized and unorganized employers and employees alike, they are also quite peculiar institutions. The labor court has exclusive competence to adjudicate labor disputes and its decisions are final and not subject to appeal. In Sweden, when hearing a case the court is normally populated by seven members, three state officials plus two members representing the employers and two representing unions.¹⁷ In Denmark, when hearing a dispute the court normally consists of seven members, three chosen by each employer and union federation, and one presiding judge, whose members are chosen jointly by the lay judges.¹⁸ These arrangements — no appeal, and direct representation of the “social partners” — keeps the regulation of labor relations insulated from the administrative apparatus of the state and the rest of the court system. This is a strikingly different arrangement than in the US, where labor relations are presided over by an administrative agency, the National Labor Relations Board, which exercises both executive and judicial functions, whose decisions are in turn subject to appeal to the federal court system, and in which labor unions have no formal representation.

Displacement

On the European continent, trade union organization falls somewhere between the two poles of Anglophone and Nordic countries. France and Germany can be used as representative cases, where striking similarities along some dimensions more or less conceal equally dramatic differences along others. In both cases, the timing of economic development imparted to trade union movements a more centralized, industrial basis with very limited workplace-based presence. Also in both countries, deep ideological divisions, both political and confessional, thwarted the organizational coherence of trade unions and led to prolonged conflict both between capital and labor up to and through World War i and ii. These facts ensured that the regulation of labor relations would assume legal and highly statist forms. However, the modes of regulation are rather different than in the Anglophone case. Given the workplace-based presence of unions, there was only one “channel” of representation in the Anglophone countries. Legal regulation in these cases therefore took the form of an intertwining dependency. By contrast, in France and Germany, where unions operated primarily at the industry level, state intervention created a second, or “dual,” channel of representation at the workplace level. This led to the displacement of labor unions.

In France, capitalism came later than in England, but also earlier and slower than other parts of Europe. Partly as a consequence of late development, the major labor federations organized themselves along industrial lines, leading to a “general absence of in-plant bargaining.”¹⁹ Consequently, unions are more inclusive than in Anglophone countries, yet at the same time more fragmented than in Scandinavia, owing to the existence of multiple confederations. These separations are the result of long-running ideological divisions within the French labor movement, where syndicalist-anarchist, communist, socialist, and Catholic currents have all vied for power and influence among the working class. It is also often observed that “the relationship between capital and labour, and among unions, is more adversarial and ideologically charged than in most other European countries.”²⁰ Another element is a strong etatist tradition of the role of government in society.

The consequence of these elements has been a French labor movement that is weak both at the workplace level (unlike Anglophone unions) and in terms of coordinating capacity (like Anglophone unions, but unlike Scandinavian unions). Despite its fame, the militancy of the ConfédérationGénéraleduTravail (CGT), earlier a syndicalist- and later a Communist-dominated organization, has been as much an impediment as a benefit to the material and organizational gains of the working class. “In reality … ‘leurgrandeetpuissanteCGT,’ as militants are somewhat romantically prone to describe it, is neither great nor powerful,” and it is in “this dissonance between words and reality [that] lies something essential to the understanding of the French Labour Movement.”²¹ Thus, while strikes were historically more frequent, collective bargaining is weak and union density is vanishingly low. The state has therefore been left to fill this regulatory void. State intervention, in fact, takes a dizzying, Kafkaesque array of forms and, given the absence of workplace-level organization, has had the effect of displacing unions more than simply rendering them dependent on the state, as in Anglophone countries.²² An example is the long-standing enterprise committees (comitésd’enterprise). In firms of fifty or more employees, law requires the election of enterprise committees by all employees (rather than, say, union members). These committees are entitled to rights of information and consultation (to “be heard,” or offer advice) on the firm’s organization, management, and general functioning, but are empowered to negotiate only on profit-sharing and financial participation.²³ These government-mandated organizations provide functions that labor unions would otherwise serve, as in the United States. Since these are rough equivalents to German works councils, we will say more about their effects below.

Germany shares certain continental similarities with France, but equally substantial differences. Capitalist development came even later than in France, but was much swifter. Industrialization did not begin in earnest until the defeat of France in the Franco-Prussian conflict of 1871, but, once started, its rapid pace allowed Germany to soon “outdistance all of its rivals” on the Continent.²⁴ Just as in France, one also cannot ignore the radicalism and internal divisions of the German workers’ movement. The Berlin general strike of November 9, 1918 led to the abdication of the Kaiser and the establishment of revolutionary workers’ and soldiers’ councils. Also prior to World War ii, “socialist, Christian, liberal and later communist and nationalist currents competed for worker allegiance.”²⁵

This history had lasting effects for the German labor movement. On the one hand, the experience of class conflict and organizational division when facing the threat of fascism compelled postwar German union leaders to form a cohesive and unitary labor union confederation, the Deutscher Gewerkschaftsbund (DGB). This gave the German labor movement far more coordinating capacity than its French counterpart. Collective bargaining has been stable and effective, granting a measure of autonomy from the state (termed Tarifautonomie in Germany).²⁶ On the other hand, a history of prolonged labor conflict has, like France, led to a legal regime of union displacement. Like France, late development ensured an absence of strong shop-floor representation and, with the memories of conflict still strong in everyone’s mind, in 1952 Germany established its famed system of works councils with legislation (Works Constitution Act) modeled on the Weimar-period Works Council Act of 1920.²⁷

A works council is a representative workplace body, whose members are elected by all workers in the establishment, not just union members. Works councils are dependent on the firm for finances and facilities and are prohibited from raising their own independent sources of revenue. Works councils are also forbidden from calling a strike or taking other industrial action, although members of the works council may participate in legal strikes called by the union. Works councils have three kinds of rights: codetermination (the right to bargain), veto, and consultation (the right to information and to be heard). The areas of codetermination are exactly prescribed and include matters such as work rules, breaks, methods of remuneration, holidays, health and safety, and the form and administration of fringe benefits.²⁸ Despite being formally representative of workers, works councils are, given this structure, sharply subordinated to the firms in which they are institutionally delimited. Thus, the works council “is situated between the economic interests of the employer and those of the employees, although it is of course nearer to the employer interests.”²⁹

There is no mystery about the motives behind the works council legislation. The Works Constitution Act of 1952 “was enacted by a conservative government with the goal of keeping unions off the shop floor and limiting their influence to sectoral bargaining”³⁰ and counteracting “a politically united union movement … that advocated nationalization of basic industries and full co-determination in line with their anticapitalist” program.³¹ The legislation itself states that “the employer and the Works Council shall work together in a spirit of mutual trust … for the good of the undertaking and its employees, having regard to the interests of the community.”³² Although unions have now accommodated themselves to the works councils regime and, in fact, have “used and adjusted them in line with their goals,” it nevertheless remains true that “works councils are now the pivotal institution of the German industrial relations system, their position vis-à-vis the union having been continually strengthened.”³³

Labor Law in the United States

The intent of this brief comparative history is to reveal the uniqueness of the form of labor union organization found in the US. Unlike either the continental or Nordic variants, labor union organization in the US (and other Anglophone countries) is characterized by strong workplace-based organization (when and where it exists) and weak coordinating capacity above the workplace level (i.e., sectoral, national, etc.). This section will trace how that form of union organization gave rise to a law-based, statist form of labor-relations regulation.

The shift to a law-based form of regulation was dramatic. Toward the end of the nineteenth century, neither unions nor collective bargaining had any legal existence. The only means available to a union to obtain recognition from an employer, bring the employer to the bargaining table, make a collective agreement, or even enforce a collective agreement, was through “extralegal” economic compulsion — the threat or exercise of strikes, boycotts, and other forms of concerted activity. Court injunctions frequently repressed such tactics — thus “recognizing” collective worker activity only in the negative sense. By the middle of the twentieth century, this had all changed: statutes established comprehensive legal regulation of all stages of a collective bargaining process presided over by an administrative agency, the NLRB, and the federal courts.

What explains this transformation? How did the regulation of labor relations come to assume the form of law? Did alternative possibilities exist?

Decentralized Unionism and the Adoption of the Legal Form

The answer I offer is that this statist regime of labor law is a product of the narrowness of labor relations themselves. Unions in the US have a strong workplace presence but weak coordinating capacity. This decentralized model of trade union organization produced pervasive employer-union conflict as well as union-union conflict. Owing to their lack of coordinating capacity, unions in the US were unable to forge a regime of self-regulation. A statist regime of labor law was constructed to fill the regulatory void.

At the heart of the 1935 National Labor Relations Act (or Wagner Act, after its main sponsor Senator Robert F. Wagner of New York) is an election procedure in which the NLRB supervises a secret ballot election and, by majority rule, awards “exclusive representation” status to a union if it prevails. Other features of the Act fit neatly into this “recognition” framework. The Act bans “unfair labor practices” to ensure that the workers’ choice of representative (or whether to be represented) is “fair and free.” After a union is “certified” by the government, the Act provides for elaborate procedures for when workers may decertify a union or an employer withdraw recognition. The legal status of various kinds of economic weapons to which workers may resort often depend on whether a union has been certified. And certification grants to unions themselves certain rights and protections, including machinery for the enforcement of union-negotiated contracts. This regime can only be described as a highly statist form of labor-relations regulation.

The origins of this majority-rule recognition procedure can be traced to the pre-New Deal era, specifically to attempts to regulate labor relations on the railroads. Union organization on the railroads is a classic example of the early-industrialization problem. First as fraternal and benefit societies, later as bona fide unions, there were no fewer than twenty different labor organizations representing workers in the railway industry. Each of these organizations, in structure and strategy, enacted the principle of exclusivity described in the previous section. “Each brotherhood, as was customary among American craft unions, claimed sole jurisdiction over the employment conditions governing employees in that craft,” whether or not the worker was a member of the union.³⁴ At approximately the same time, railway unions began appealing to the majority-rule principle both to justify their demands for union recognition vis-à-vis employers and to solve their jurisdictional disputes with one another. This all took place against the backdrop of extraordinary labor strife. Later, this principle was adopted in one of first pieces of national legislation regulating labor relations, the Transportation Act of 1920. Fifteen years later, a series of statutes, court decisions, and policy choices had so narrowed the available options that “the question of Wagner’s intent became secondary to his policy constraints. Wagner built the NLRA upon an ideology that had become self-sustaining.”³⁵

Scholars have criticized the NLRA for enshrining into law the old AFL’s “voluntarist” labor-relations philosophy. This was accomplished either by the passage of the NLRA itself or by its subsequent “judicial deradicalization.” Either version treats the NLRA as a kind of ex nihilo event, without any legal or policy history of its own.³⁶ Ruth O’Brien convincingly demolishes this account. It was not the AFL’s voluntarism that prevailed but the progressive movement’s “responsible unionism.” For progressives, the labor movement was too narrowly self-interested to accommodate the “public interest.” What was needed was a Hobbesian strong state — one that would subordinate the labor movement to the “true” guardian of the public interest.³⁷ I endorse O’Brien’s version of events, but she doesn’t account for the counterfactual: could the AFL’s voluntarism have been a viable alternative solution to the “labor problem”? Given the lack of coordinating capacity among US labor unions, I suggest not. At least partly, the progressives’ critique of the AFL-dominated labor movement was true. It is just that the possibilities, if not the concrete choices available to the labor movement in the early 1900s, were not limited to either a Leviathan or narrow craft voluntarism. The following comparative example makes this claim concrete.

In a forgotten story in labor history — forgotten because of the opportunity that was not taken — the International Association of Machinists (IAM) and the National Metal Trade Association (NMTA) signed the so-called Murray Hill agreement in 1900. In terms of the agreement’s substance, employers conceded to a reduction in the working day from ten to nine hours for all machinists in NMTA shops. However, a complication arose from the union’s inability to convince all NMTA employers to also adopt a uniform 12.5 percent wage increase to maintain weekly earnings at earlier levels. The agreement was repudiated in the following strike wave, the union claiming that the employer had failed to agree to the wage increase, the employers accusing the union of calling strikes instead of settling the disputes through the central arbitration system established by the agreement. As told by Peter Swenson, employers would have in time accepted, and many would have even welcomed, centralized bargaining over wages and working conditions in exchange for the unions relinquishing their job-control objectives. Employers “slammed the door shut for all time, however, because union militants used the strikes to impose the closed shop … and rules prohibiting men from operating more than one machine at a time, working for piece rates, and instructing unskilled workers.”³⁸ The IAM leadership did not approve the strikes and in fact had agreed to management’s demand for the open shop and the right to manage. Thus, the objective of taking wages out of competition came to founder on the IAM’s inability to control local militancy and designs on job control.

At almost exactly the same time, in 1905, an almost identical experiment in the identical industry led the Swedish labor movement in a very different direction. Confronted with a metal-workers’ strike, the employers’ association in the engineering industry responded with a lockout at eighty-three member firms. The conflict led to the “first industry-wide multi-employer wage settlement for any industry in the country.” The agreement “allowed no restrictions on manning of machinery or hiring of unskilled workers and apprentices … [and] the union agreed to an open shop clause.” The metal workers’ counterpart in the United States, “[m]ilitant skilled craftsmen” in the IAM, “would have regarded the deal with dismay and disgust.” The next year, this industry agreement was followed by a multi-industry, national agreement known as the “December Compromise.” A key section of the agreement prohibits closed-shop agreements and establishes management control over “decisions involving hiring, firing, and supervising work.”³⁹ Yet what workers gave up in firm-level “production politics” they gained in power over the labor market itself. Centralized bargaining has come to deliver high union density, the lowest level of wage dispersion in the advanced capitalist world, and most critically, high inclusivity, encompassing virtually all wage earners.

The IAM’s attempt at establishing industry-wide bargaining vividly demonstrates how the US labor movement’s workplace-centered unionism acted as an obstacle to broader and more inclusive forms of worker organization. Centered at the workplace, and pursuing a job-control strategy, US unions had significant power to contest the employer’s domination of the labor process. Unfortunately, for exactly those same reasons, this constellation of power was too weak, too uncoordinated between firms, to contest the domination of the market. As the comparison of the IAM with the Swedish metal workers shows, local power generated conflict but obstructed efforts to develop self-regulation. Following decades of the “labor problem,” the state stepped in as regulator. As a result, “[g]overned by this state-operated regulatory agency [i.e., the NLRB], organized labor no longer shaped its own destiny—it was dependent on this agency.”⁴⁰ O’Brien is therefore correct to insist that it was the progressives’ statist vision rather than the AFL’s voluntarist philosophy that prevailed. Nevertheless, we should not overlook how historically given forms of labor organization frustrated other possible forms of labor-relations regulation. This gives us another reason why voluntarism per se was not the culprit in labor’s current legal and existential crisis. Scandinavian self-regulation is, after all, another kind of voluntarism. At the same time, as the IAM example demonstrates, the institutional and organizational narrowness of craft unionism left the door open to a statist regime of labor law.

The Consequences of Ignoring the Legal Form

Because of unions’ strong workplace presence but weak capacity for coordinating activity across workplaces, the regulation of labor relations was achieved by recourse to the law. This claim cuts directly against the thrust of a tradition of “critical” labor law. The story told by critical labor law scholars is of a potentially “anticapitalist” National Labor Relations Act that was “deradicalized” by conservative judges and narrow-minded intellectuals.⁴¹ In these approaches there is never any question whether the law should be used to regulate labor relations. Rather, the line of attack is to challenge the particular content of the labor law, not the form of regulation itself. Not only is this a mistake as a method of analysis but, as I will also demonstrate, it also commits an instrumentalist error about the nature of the law and the state within capitalism.

A content critique of law obscures the way that law does more than simply help or hinder the labor movement achieve various, specific objectives. As a form of social regulation, the law also allocates determinate material and ideological resources as a means to achieve these ends. These means threaten to substitute for the working class’s own material and ideological means of regulation. This would not be an issue if labor unions or other working-class organizations were merely means of achieving gains for workers. But they are not. Whatever their limitations, unions are moments in the process by which workers constitute themselves as a class. Thus, the law — not in its content, but as a form of social regulation — always presents the danger of undermining this process through mechanisms of dependency and displacement.

Let me demonstrate these problems with a few examples. First, left and liberal criticism of labor law has duly taken note of the stunning raft of restrictions on worker economic action under the NLRA. One of these is the prohibition on picketing an employer when the objective is to get the employer “to recognize or bargain” with a labor union, unless the union is already “currently certified” by the NLRB as the employees’ representative. Falling under this ban are not only pickets where a well-known and well-funded labor union demands recognition from an employer; the ban also covers pickets where workers themselves have formed their own, independent union, and demand either recognition or, more simply, a raise in wages, shorter hours, or any other change in the terms and conditions of employment.

The practical — or what I will describe as the sociological — effect of this ban is to compel workers and unions, in lieu of concerted activity, to instead make use of the Board’s election and certification procedures, which are slow, cumbersome, and allow the insertion of the employer, via its free speech rights, into the union-formation process. As research has shown, union success in certification elections began to fall once “employers gained significant access to employee decision-making in representation proceedings in the mid-1940s ….”⁴² The Board’s election procedure transforms what would otherwise be an organic process of worker self-organization into a mechanical (and legalistic) one where, ultimately, the formation of a union depends on the singular, individual act of voting. Undoubtedly, this amplifies the ubiquitous employer rhetoric that fashions unions as “third parties” intent on upsetting the intimate decorum of the familial (or, should we say, paternalistic) employer-employee relationship.

While I agree, emphatically, with left critics of labor law, the mistake is to interpret this sorry state of affairs as the result of ideologically motivated judicial decision-making. The ban on recognition or bargaining picketing is not intended to quash worker self-organization. Rather, it exists because the law already provides a legally protected right to be represented by a union, a right made effective through a “fair,” “independent,” and “neutral,” government-supervised election procedure. Viewed through the lens of the law itself, recognition and bargaining picketing is incongruous, even intrinsically lawless insofar as it undermines existing, legally ordained procedures. Indeed, although this picketing ban was added by the anti-union Taft-Hartley Act, the Board had already withdrawn protection for strikes “to compel an employer to recognize or bargain with a union other than that certified by the Board” before Taft-Hartley’s passage.⁴³

Even invoking the category of “ideological” to explain distortions in the Board’s election procedure runs into difficulties. However laudable the goal that employers should also remain neutral in the election process — employees’ decision to form a union should be theirs alone — “legally, the Board clearly found the imposition of strict neutrality difficult to defend against free speech claims advanced in the same discourse of legalities that it had chosen as its language.”⁴⁴ Lawyers, labor law administrators, and especially federal judges don’t spend their time thinking about the sociological implications of legal rules nor the complexities of class formation, nor are they really equipped to do so. They do think about accommodating conflicting rights when presented with claims between formally equal, distinct and private, legal personalities within an adversarial process. This too, of course, is “ideological” but it is intrinsic to, even identical with, the legal form itself, not an extrinsic, corrupting influence to some supposed “anticapitalist” legislation.

A second recurring target for left critics of labor law is the ban on strikes during the life of a collective agreement. Not only are strikes banned, but also an employer may turn to a federal court and have the strike immediately enjoined, rather than wait for a damage award following a lengthy trial. On first appearance, this seems like an open-and-shut case: not only is the ban particularly harsh in its method of enforcement, but also as a direct prohibition on worker self-activity, it “clearly” reflects “a repressive, antilabor spirit,” according to labor law scholar Karl Klare.⁴⁵

A closer look reveals some problems with this perspective. First, as many of these same critics are forced to acknowledge, labor “peace” can often also work to the advantage of workers.⁴⁶ Why should an employer grant wage concessions, or for that matter ever concede to an agreement, let alone tolerate the existence of a union, if the union doesn’t “uphold its end of the bargain”? Furthermore, courts will enforce the strike “ban” only where the union has explicitly agreed not to strike during the term of the agreement or when the union has agreed to arbitrate disputes under the contract (implying, according to the courts, an agreement not to strike). Moreover, federal labor law explicitly empowers federal courts to enforce these agreements to arbitrate on unions’ behalf. Indeed, the Supreme Court decided that injunctive relief was available to an employer to enforce a no-strike agreement in part because unions had been able to obtain similar remedies to enforce their rights under the contract.⁴⁷

Labor-relations history in France provides an instructive example here. In law there is no prohibition of strikes during the term of a collective agreement and unions as a matter of principle have refused to agree to “no-strike” or “peace” clauses. But, if anything, such unrestrained strike power has contributed to the inability of unions to gain much through collective bargaining. Instead, as illustrated in the previous section, it is the state that has assumed the central role in governing working conditions and workplace representation.

Most critically, US law already provides for enforcement of labor contracts — strikes are legally redundant. Writing for the majority, Justice Brennan stated, “Indeed, the very purpose of arbitration procedures is to provide a mechanism for the expeditious settlement of industrial disputes without resort to strikes, lockouts, or other self-help measures. This basic purpose is obviously largely undercut if there is no immediate, effective remedy for those very tactics that arbitration is designed to obviate.”⁴⁸ Of course, sociologically, self-help and arbitration are not redundant: if strikes are the means by which workers build solidarity and develop class consciousness, then the substitution of the strike with the other means of reaching working-class objectives may — intentionally or unintentionally — undermine working-class interests. But, again, the point is that the existence of a legal right itself entails a restriction on the ability to strike to protect the same interest.

A third example concerns the legal status of concerted activity taken in response to an employer’s unfair labor practices. The Supreme Court addressed this issue in a widely cited and discussed decision, NLRB v. Fansteel Metallurgical Corp.⁴⁹ In that case, the employees responded to a series of the employer’s unfair labor practices — recognizing only an “independent,” company-dominated union, and employing a labor spy to engage in espionage within the bona fide, “outside” union — by “seizing the employer’s property” in a sit-down strike. The employer countered by announcing that “all of the men in the plant were discharged for the seizure and retention of the buildings.” The employer then appealed to the local sheriff, who with an “increased force of deputies” evicted the workers from the plant and arrested them; most of the workers were eventually fined and given jail sentences. As a remedy for the employer’s unfair labor practices, the Board ordered “‘immediate and full reinstatement to their former positions,’ with back pay.” However, the Supreme Court denied enforcement of this order, concluding that the workers had been legitimately discharged for illegally seizing the employer’s property.

The court’s decision has been widely criticized for taking a narrow view of “concerted, protected activity,” and ignoring the workers’ claims to be acting in self-defense against the employer’s violation of their rights granted to them by the Wagner Act. According to Karl Klare, the language of the Fansteel decision reinforces the role of workers as sellers of labor power and consumers of commodities, rather than as producers, and obstructs an alternative perspective presaged by the “‘dereifying’ explosion of repressed human spirit” expressed in the sit-down strike.⁵⁰ According to James Gray Pope, the Fansteel decision inverts appropriate legal hierarchies, placing the employer’s common-law property rights above those of the employee’s statutory right to engage in collective action, a conclusion that can only be justified by an unstated appeal to a discredited interpretation of the Constitution.⁵¹

Both critics, however, overlook the very first words of Chief Justice Hughes’s decision following its statement of facts: “For the unfair labor practices of [the employer] the Act provided a remedy. Interference in the summer and fall of 1936 with the right of self-organization could at once have been the subject of complaint to the Board.”⁵² Once again, using the strike to enforce workers’ statutory rights is legally duplicitous because the Board already possesses the power to enforce those rights. The court continued, “To justify such conduct because of the existence … of an unfair labor practice would be to put a premium on resort to force instead of legal remedies and to subvert the principles of law and order which lie at the foundations.”⁵³ Responding to this language, Klare is correct to draw attention to the inherently peaceful nature of workers’ concerted activity in general and the sit-down strike in particular.⁵⁴ But it is not the court’s hysterics that are most interesting; instead, it is the overlooked rationale that, whether violent or not, concerted action to enforce rights already subject to Board administration and enforcement subverts the appropriate scheme of rights enacted by the NLRA. Thus, it is not (or not just) ideologically freighted judicial reasoning that has undermined the labor movement, but the very rights themselves, created and enforced by the state apparatus, that have justified restrictions on concerted worker activity.

We might be content simply to revise the ideological critique of labor law to include the insight that there is “something integral to the law itself, and its purposes” that has contributed to the weakness of the contemporary labor movement.⁵⁵ But the ideological critique of labor law also has prescriptive drawbacks for labor strategy. As I will now show, an analysis of these critiques reveals an implicit and problematic view of the nature of the state and law. This view is an instrumentalist one, not in the sense of “a conspiratorial model of political and legal decision making, in which all outcomes consciously serve the interests of identifiable business elites,”⁵⁶ but rather in its “tendency to assume that the state as an instrument … can be used with equal facility and equal effectiveness by any class or social force.”⁵⁷ This is because these critics focus on the content of labor law rather than the form of regulating labor relations. Implicit in that kind of critique is an acceptance that law is the appropriate form of regulation.

Let me illustrate this claim with a couple of examples. An almost universal complaint is the allegation of “voluntarism” as the defining feature of US labor law. We have already rejected this as the defining feature of US labor law, but it certainly has its place. Thus US labor law lacks any substantive criteria for regulating the content of collective agreements made by unions and employers. As Karl Klare describes it, voluntarism is simply the importation of a broader idea of contractualism, necessary for a capitalist economic order, into the collective bargaining relationship:

The central moral ideal of contractualism was and is that justice consists in enforcing the agreement of the parties so long as they have capacity and have had a proper opportunity to bargain for terms satisfactory to each. Contractual justice is, therefore, formal and abstract: within the broad scope of legal bargains it is disinterested in the substantive content of the parties’ arrangements.⁵⁸

According to Klare, the Supreme Court’s landmark decision in NLRB v. Jones and Laughlin Steel Corp. determined that “the wage-bargain would remain fundamentally within the contractualist, private ordering framework” of contract.⁵⁹ In that decision, Chief Justice Hughes wrote that the Act “does not compel any agreement whatever,” but leaves the making of agreements and their contents to the parties themselves.⁶⁰ Therefore, the Supreme Court rejected, in Klare’s view, the choice of “state regulation of the substantive terms of the wage bargain.”⁶¹

Hold that thought for a moment. Later, Klare examines what he calls the “public right doctrine.” In the Supreme Court decision, American Federation of Labor v. NLRB, the majority held that an employer could not challenge in federal court the Board’s decisions in representation proceedings, such as the determination of the bargaining unit, “unless and until the employer was charged with an unfair labor practice” and raised a bargaining-unit issue as a defense. According to Klare, “This decision implicitly rested on the idea that the statutory scheme does not protect private entitlements but protects certain public interests.” Although initially seen as a victory for labor, that assessment was short-lived. If the party aggrieved in a representation proceeding was the union, and the union lost an election because of the definition of the bargaining unit, it would have no remedy under the Act, because it would be impossible to file a refusal-to-bargain unfair labor practice — because the union lost the election, the employer would not be obligated to bargain. Klare attacks these and other decisions for creating an intellectual justification for the dependency of labor on the state, thereby reinforcing the cultural hegemony of liberal political theory. This dependence hindered labor from conceiving itself, or acting, as an autonomous movement capable of fundamentally transforming the established social relations of production.⁶²

One can only agree with Klare’s diagnosis of dependency — I have argued that this is a hallmark feature in Anglophone labor law. But it is difficult to reconcile Klare’s assessment in this case with his earlier conclusion in Jones & Laughlin. In Jones & Laughlin, he criticizes the Court for embracing private-ordering, for not inquiring into the substantive terms of collective agreements, for not remedying any power imbalances, for not licensing state regulation of the wage bargain. But later, when interrogating the public right doctrine, he complains about the state determination of bargaining units and even questions why the Board should have “the power to define the balance of opposing economic forces on which the substantive outcome of collective bargaining depends” and thereby the power to delimit the “contours of legitimate class struggle.”⁶³ In other words, Klare criticizes the Court in Jones & Laughlin for not substantively regulating labor relations; but then he criticizes the court in American Federation of Labor for doing exactly that. Would not “substantive regulation of the wage bargain,” for precisely the same reasons, prevent the labor movement from acting as an “autonomous movement capable of fundamentally transforming the established social relations of production”?⁶⁴ As we saw in the case of Australia, there is good evidence for precisely that expectation. We can agree with Klare about the dangers of dependency, but his content-centered critique of the law prevents him from consistently carrying this argument through. This or that legal rule may or may not favor workers’ immediate objectives, but the law, as a form of social regulation, may undermine working-class self-organization because such self-organization is not simply a means toward achieving working-class objectives but is itself an alternative form of social regulation.

A second example of instrumentalism comes from the labor law rule allowing employers to permanently replace workers striking for wages or other economic objectives. The origin of this rule can also be traced to an early Supreme Court decision.⁶⁵ As several labor law scholars contend, the employer’s ability to permanently replace workers has had a devastating impact on the labor movement. These consequences have led to habitual calls to ban the use of permanent replacements, either through legislation, as was unsuccessfully attempted in the early 1990s,⁶⁶ or through case-law adjudication, on the theory that the Supreme Court’s decision was wrongly decided and is inconsistent with the language and policy of the NLRA.⁶⁷

Suppose that the US Congress was to pass legislation or the Supreme Court to overrule precedent banning the use of permanent replacements during a strike. Why stop with permanent replacements? If the use of permanent replacements is harmful to the labor movement, certainly the use of temporary replacements is as well? If temporary as well as permanent replacements were banned, workers’ bargaining power would certainly be strengthened. But, of course, it is not really workers’ collective activity that is the source of the bargaining power — rather, it is the government that compels the employer not to hire replacements. And if we agree that it is acceptable, or even better, to substitute state power for workers’ power in this case, why limit the use of state power just to strikes? Why not simply let the state, rather than unions, set wages and working conditions? Why bother with unions at all? In the concluding section, I will offer my own solution to this problem. But by now the point should be clear that labor law critics have yet to offer any rationale, justification, or guiding principle whatsoever about when to substitute law for worker collective action. I am not claiming that some justification, however rough or fine, cannot be found. Rather, I am claiming that, because their analysis is limited to a critique of the content of labor law, they have ignored the ways that the law, as a form of regulation, can unwittingly undermine those essential processes of class formation for the sake of immediate objectives. If these dangers are present for working-class formation in a way that they are not for capitalists, it should also be clear why both these examples betray an instrumentalist view of the state and law.

What then should be the attitude of the labor movement toward the law? The very existence of the state and law requires some engagement with it, if only to avoid it. I address these issues in the next section.

Labor Law and Union Strategy

I have argued that the regulation of labor relations need not always assume the form of law, and that in fact it does not always assume the extreme form of legalism that we find in the United States. I have also demonstrated the contradictory nature of rights in the regulation of labor relations. What kind of labor legal strategy emerges from this analysis?

The introduction drew the distinction between rights and freedoms.⁶⁸ Rights are those interests or actions that are protected by the coercive power of the state. Freedoms on the other hand are those interests or actions that are not prohibited by the state, but also with which others may interfere; freedoms are neither legally protected nor prohibited. My contention is that the labor movement should advance labor freedoms and be wary about labor rights.

This contention follows from the previous analysis. Since rights are distinguished by the fact that they are protected by the coercive power of the state, bureaucrats, judges, and legislators can use that fact to restrict labor’s own means and powers to enforce these interests and claims, subordinating society to the state. Indeed, as I have shown, state officials, with interests and power of their own, are likely to view labor’s competing power as legally redundant and particularly subversive. Labor freedoms restrict the coercive power of the state in a way that gives priority to labor’s autonomous sources of power, subordinating the state to society.

Advancing labor freedoms is hardly an unambitious strategy, since direct prohibitions on concerted activities are abundant. The three most restrictive prohibitions on strike activity are those directed to (1) mass picketing,⁶⁹ (2) organizing and bargaining strikes,⁷⁰ and (3) secondary strikes and boycotts.⁷¹ Each is an affirmative ban on worker collective action, by which an employer may have the actions enjoined and the union fined. As such, they are restraints on workers’ freedom of action. The first ban has done the most to destroy the power of the strike and, as discussed below, to open the door to the employer’s use of replacement workers. The second has done the most to squelch coordinated worker activity across firms and industries. As identified earlier, the third has done the most to derail and suppress organic worker self-organization. These restrictions could be eliminated through various means. Congress could amend the National Labor Relations Act, and remove the offending provisions. Some labor law scholars have argued that these provisions violate the First Amendment and therefore should be declared unconstitutional. The labor movement should entertain all options, but I have little doubt that massive civil disobedience though direct worker confrontation with these legal barriers will also be necessary to discredit and overcome them.

If such labor freedoms were achieved, employers would be under no state-imposed duty to refrain from interfering with workers engaged in such activities. Workers could be terminated for engaging in mass picketing, organizing strikes, or secondary picketing. Freedoms may therefore strike some readers as insufficient. Yet, it has been the burden of this essay’s comparative, historical, and legal analysis to demonstrate the self-defeating sociological effects of labor rights. Nevertheless, there is truth to the claim that certain, fundamental labor rights remain essential. Thus, insofar as it facilities worker solidarity and collective action, there seems little reason to eschew, for example, a worker’s right to join a union. Even more fundamentally, the rights of workers to be free from the employer’s physical assaults or from the state’s interference with speech and expression are also necessary. The distinction between rights and freedoms is no talisman. Rather, the ultimate objective must be kept in mind: the collective self-organization of the working class.⁷²

To convince the reader that this proposal is not merely wishful thinking, we should recall the self-regulation models of Scandinavia. In Denmark and Sweden, the regulation of labor relations — including such fundamental matters as union recognition and minimum wages — falls within the purview of unions and organized employer associations. Strikes that are banned in the United States remain viable options in Scandinavia. Enforcement of the rules and agreements depends primarily (though not exclusively) on the economic weapons of labor and employers, rather than the physical compulsion administered by the state. Labor courts, unlike the NLRB, operate outside the hierarchy of the bureaucracy and courts of the state apparatus.

While the Scandinavian experience demonstrates the viability of a regime of labor freedoms, one may also raise the objection of Nordic exceptionalism. There are several responses to this objection. In the main, however, my proposal does not call for a wholesale replication of the Nordic model. In fact, that model ultimately demonstrates the inherent limitations of social democracy. Instead, the self-regulation model establishes the viability of governing the labor market through workers’ own organizations, autonomous from the state. This experience will be absolutely vital, whatever specific path the revitalization of labor takes — and not only for rebuilding the labor movement, but also during a period of a “lower-stage of communism,” where labor is still compensated according to differential ability, but becomes increasingly independent of the competitive labor market and market criteria. There is no reason why we cannot learn, replicate, or even build on particular features of this history.

This proposal to favor labor freedoms over labor rights also implies some skepticism toward other proposals for labor law reform. Richard D. Kahlenberg and Moshe Marvit double-down on a labor-rights strategy in their Why Labor Organizing Should Be a Civil Right.⁷³ James Gray Pope, Ed Bruno, and Peter Kellman have also advanced a rights-centered proposal in the Boston Review.⁷⁴ Alex Gourevitch has recently made a compelling case specifically for the “right to strike.”⁷⁵ My concerns with each of these proposals should be clear by now.

Focusing on Gourevitch’s proposal is instructive. In order to restore the power of the strike, his proposal would prohibit an employer from hiring permanent replacement workers during a strike. Above, I raised some hesitations about this idea. But there is another solution to the problem of permanent replacements, one that depends directly on worker solidarity and of which labor once made extensive use. Although the Supreme Court announced the rule allowing permanent replacements in 1938, in NLRB v. Mackay Radio & Telegraph Co., employers initially made little use of it. The reason was that in the period of labor upsurge in the 1930s and 1940s, mass picket lines effectively deterred the hiring of temporary, let alone permanent, replacement workers. These actions contributed to an “unofficial norm” that prevented employers from hiring replacement workers, although the official law permitted them to do so.⁷⁶ Therefore, from the standpoint of labor freedoms, the real problem with permanent replacements is the decline in the unofficial norm that prevented the police from interfering with picket lines and dissuaded employers from hiring strikebreakers. And on this score the prohibition and erosion of workers’ freedom to engage in mass picketing — the activities that established the unofficial norm in the first place — were paramount.⁷⁷ The legal solution to labor’s problems should then be directed to rolling back the prohibition on mass picketing, rather than to the creation of a right to reinstatement immediately after a strike.

Finally, my proposal should be distinguished from a kind of myopic syndicalism that can be found on the radical labor left. On the one hand, this essay shares much in common with the syndicalist approach, particularly on the need to reengage mass worker collective action. There is, however, a key difference between self-regulated and unregulated labor relations. This essay’s proposal falls into the former category, syndicalism into the latter. Syndicalism’s view eschews any restriction on concerted activity — even in cases where unions voluntarily agree not to strike during the term of the collective agreement.

Another important difference is that syndicalism frequently overlooks the organizational dimensions of union power, both ignoring the decentralization and lack of coordinating capacity that, as I have argued, are serious problems for labor in the US, and chaffing against union and bargaining centralization found in places like Sweden.⁷⁸ In terms of actually accomplishing gains for workers, however, it is hard to ignore the Nordic experience. At the same time, the glaring shortcomings of the militancy-above-all-else syndicalism in France and the workplace-centered, decentralized bargaining in the US are all too real. From my perspective the organizational question is at least as important one as the legal question, but lack of space requires me to defer this topic for later discussion.

A final point of distinction to make is that my perspective hinges on the specific historical conjuncture. Currently, in the United States, a working class barely exists — in the class-for-itself sense. We are in a war of position, which requires that we develop a long-term strategy to build the strongest, most cohesive working class possible. That is, we need to build labor unions’ capacity for autonomous, class-wide action. If a war of maneuver emerges at some future point, perhaps a primary emphasis on militancy will be required. The kind of political movement and organization that unions are allied with in these circumstances, rather than unions’ specific organizational form, will then become the most decisive issue. But until then, our project should be a counter-hegemonic one.

Conclusion

This essay has sought to demonstrate three distinct points. First, the institutional history of unions in the US is unique. In particular, the timing of unions’ development within capitalism and their lack of coordinating capacity set them apart from unions on the Continent and in Northern Europe. Second, the organizational weakness of unions is an important factor explaining why the regulation of labor relations in the US assumed such a highly statist and legal form. Third, this legalistic regime of regulation has been inimical to the labor movement. Judges, bureaucrats, and legislators have often cited statutorily conferred rights as reasons for restricting workers’ concerted activities. Just as damaging, these same, negative effects of labor rights have been obscured by critics who have focused on the content of labor law rather than the legal form itself. As a strategy to overcome this history, I offer the legal right/freedom distinction as an organizing principle that can guide the labor movement’s attitude toward the law. It should not serve as an inviolable distinction, much less one that assumes an independent life of its own. Nevertheless, it is, I submit, a compelling way to capture Marx’s idea of law and politics: the idea of converting the state from an organization superimposed on society, to one subordinate to it, an objective inseparable from building a socialist society. Echoing this sentiment are the words penned by Christopher Tomlins at the end of his book, The State and the Unions: “[A] counterfeit liberty is the most that American workers and their organizations have been able to gain through the state. Its reality they must create for themselves.”