How Trade Unions Changed Labor Markets

The standard competitive model of labor markets treats wages as the outcome of supply and demand between many workers and many employers, each of whom is a price-taker without market power. In this model, unions are an interference with competitive equilibrium, raising wages above their market-clearing level, reducing employment, and creating an inefficient wedge between the marginal product of labor and the price at which labor is hired. This is a coherent model with real empirical support in some contexts. It is also radically incomplete as a description of the labor markets in which trade unions actually emerged and operated. The early industrial labor markets of Britain, the United States, Germany, and France were not characterized by many competing employers competing vigorously for workers. They were characterized by spatial monopoly, information asymmetry, and the catastrophic vulnerability of workers who had no savings, no alternative income, and no legal recourse. Understanding unions as economic institutions requires starting with the market failures that made them rational responses rather than with the idealized competitive benchmark from which they deviated.

The company town is the emblematic case of labor market monopsony. When a single employer owns the mine or mill around which an entire community has grown, workers have no effective alternative employer within walking distance. The coal miners of Appalachia, the mill hands of Lancashire, the steel workers of Gary, Indiana, were not choosing among competing employers when deciding whether to accept the offered wage — they were accepting or refusing the only offer available. The employer in a company town is a monopsonist in the labor market, just as a monopolist is the single seller in a product market. A monopsonist has market power over the workers he employs: he can pay them less than the competitive wage precisely because their alternatives are so limited. The wage will be set below the workers’ marginal revenue product — the value of what they produce — and the employer captures the difference as a monopsony rent. This is not an imperfection that the labor market will self-correct; it is a structural feature of the market that persists as long as the geographic isolation and employer concentration persist. Unionization in this context is not a departure from competitive equilibrium; it is a countervailing power that offsets the employer’s monopsony power and pushes wages back toward the competitive level.

Information asymmetry compounded the monopsony problem. Employers knew wages paid across the industry; workers generally did not, in an era before labor market surveys, published wage data, or union wage books. A worker negotiating individually with an employer was negotiating from profound informational disadvantage. He did not know whether the wage being offered was typical or below typical, whether other workers in the same plant were earning more, or whether the employer’s claims about the firm’s financial capacity to pay were accurate. The employer could exploit this information asymmetry systematically. One of the most important early functions of trade unions was aggregating and disseminating wage information: the union wage book, published for members and publicly available, established a common standard that reduced the employer’s informational advantage. When workers knew the union rate, they knew what their labor was worth by the industry standard and could negotiate on that basis. This information function preceded and was analytically distinct from the collective bargaining function, and it mattered greatly in markets where the information asymmetry was otherwise severe.

The legal environment of early industrial labor markets reinforced both problems. Under British common law through most of the early nineteenth century, combination — that is, workers agreeing collectively to refuse to work except at a specified wage — was a criminal conspiracy. The Combination Acts of 1799 and 1800 made explicit what common law had often implied: workers who organized collectively to raise wages could be prosecuted. The legal framework was not neutral; it actively suppressed the collective action through which workers might have offset employers’ monopsony power. The legal situation changed gradually: the Combination Acts were repealed in 1824-25, though subsequent legislation continued to restrict strike action; British unions gained full legal protection in 1871 and 1875; American unions achieved comparable protection under the Wagner Act of 1935. This legal history is part of the economic history of unions because the legal environment determined when collective organization was viable, not just legally protected but practically effective. A union whose members could be prosecuted for striking was a much weaker institution than one whose right to strike was legally secure.

The economic effects of unionization, where unions achieved genuine bargaining power, were several and not uniformly positive or negative by any single welfare criterion. The wage effect is the most studied: unionized workers in the American economy of the 1950s through 1970s earned approximately 15 to 25 percent more than comparable non-union workers. This wage premium was real and substantial, and it was not simply a reflection of unionized workers being higher-skilled — careful econometric studies controlling for observable worker characteristics consistently find a premium attributable to union membership itself. The distributional effect of this premium was to compress the wage distribution: unions raised wages at the bottom and middle of the skill distribution more than at the top, reducing within-industry wage inequality. In the high union-density decades of the mid-twentieth century, this wage compression effect was a significant force reducing overall income inequality in the United States and other unionized economies.

The employment effect of union wage premiums is theoretically predicted to be negative — if wages are above the competitive level, employment should be lower than it would otherwise be. The empirical evidence on this question is more ambiguous than the theory suggests. Richard Freeman and James Medoff’s influential study What Do Unions Do (1984) found that the negative employment effect of union wage premiums was modest and was partly offset by a positive productivity effect. Their argument was that unions, by giving workers a voice through collective grievance procedures and seniority systems, reduced labor turnover (which is costly to firms) and improved worker morale and cooperation. The exit-voice framework, drawn from Albert Hirschman, proved analytically useful: workers who would otherwise quit their jobs — exiting the employment relationship — might instead use the union’s grievance procedure to raise concerns and have them addressed, remaining more productive employees. This productivity offset partially compensated employers for higher union wages, which is one reason why unionized firms were not systematically less profitable than non-union competitors during the high union-density period.

The seniority system deserves particular attention because it was one of unions’ most economically consequential institutional innovations. In non-union workplaces of the early industrial period, employment security was entirely at the employer’s discretion: workers could be laid off for any reason or no reason, and the order of layoffs during downturns was determined by supervisors’ personal judgments. This discretion created scope for favoritism, political management of the workforce, and the use of layoff threat as a disciplinary tool that individual workers had no means of contesting. The seniority system — last hired, first fired — was a union institutional innovation that replaced this discretionary system with a transparent rule. Seniority rules reduced favoritism and arbitrary dismissal, which reduced workers’ fear of retaliation for legitimate workplace grievances and made them more willing to invest in firm-specific skills. But seniority systems also reduced employer flexibility in workforce adjustment and protected poor performers whose seniority exceeded their productivity. The net economic effect depended on which factor dominated in specific industries and contexts.

Union density in the United States peaked in the mid-1950s at roughly 35 percent of the private sector workforce and has declined continuously since, reaching below 7 percent of private sector workers by the 2020s. This decline is one of the most important economic developments of the late twentieth century, and explaining it requires separating several distinct mechanisms. Structural change in the economy shifted employment from the heavily unionized manufacturing and mining sectors to the service sector, which has historically been harder to organize. Globalization increased competitive pressure on unionized manufacturing firms, raising the cost of union wage premiums and shrinking the unionized manufacturing sector through both plant closures and offshoring. Legal changes, particularly the NLRB’s increasingly employer-favorable interpretation of the Wagner Act from the 1970s onward, raised the cost and reduced the effectiveness of union organizing campaigns. And the deliberate union-avoidance strategies of American employers — sophisticated legal tactics, anti-union consultants, and the credible threat of plant closure in response to organizing drives — became more aggressive and more effective.

The relationship between union decline and rising wage inequality in the United States after 1980 is one of the better-documented regularities in labor economics. Card, Lemieux, and Riddell have estimated that declining union density accounts for roughly one-fifth of the increase in male wage inequality between the 1970s and the 1990s — a substantial fraction, though not the majority. The mechanism is not only that unionized workers’ wages fell relative to non-union workers as union density declined, but that the existence of a large unionized sector had historically set wage standards that non-union employers also felt pressure to match. When union density was high, the threat of organizing drove non-union employers to pay closer to union wages to prevent unionization. As union density declined and the threat of organizing receded, non-union employers could pay less, and the wage floor that unionization had effectively established for the broader labor market eroded.

The international comparison illuminates what union density requires to be maintained. Scandinavian countries maintained union densities of 60 to 80 percent through the late twentieth century — far higher than the Anglo-American economies — through institutional arrangements that linked union membership to unemployment insurance administration (the Ghent system). Workers in Sweden, Denmark, and Finland had strong financial incentives to join unions because union membership provided access to unemployment benefits administered through union funds. This institutional linkage meant that union density did not depend entirely on workers’ collective action capacity against employer resistance; it was supported by the design of the welfare state itself. The contrast with the American model — in which unions must organize plant by plant against determined employer opposition, and membership provides no comparable welfare benefit linkage — explains much of the divergence in union density trajectories.

The broader lesson of union history for understanding labor market institutions is that the power of workers to bargain effectively depends on institutional conditions that do not arise spontaneously from market forces. Monopsony power, information asymmetry, legal restrictions on collective action, and employer anti-union strategies all work systematically against workers’ bargaining capacity. The periods and places where workers have successfully organized and maintained bargaining power are precisely the periods and places where these structural obstacles were reduced: by legislation that protected collective bargaining rights, by labor market tightness that increased workers’ outside options, by political coalitions that gave labor movements access to state power, and by institutional designs — like the Ghent system — that embedded union membership in the welfare state rather than leaving it entirely vulnerable to employer resistance. When these conditions deteriorated — as they did in the Anglo-American economies from the 1980s onward — union power dissipated despite workers’ continued desire for the wage and voice benefits that unions provided. The conditions for effective worker organization are themselves the product of political and institutional choices, not the inevitable output of labor market supply and demand.