1
Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and certain other countries.
Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America.
© Oxford University Press 2021
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by license, or under terms agreed with the appropriate reproduction rights organization. Inquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above.
You must not circulate this work in any other form and you must impose this same condition on any acquirer.
Library of Congress Cataloging-in-Publication Data
Names: Posner, Eric A., 1965– author.
Title: How antitrust failed workers / Eric A. Posner.
Description: New York, NY : Oxford University Press, [2021] | Includes index.
Identifiers: LCCN 2021017327 | ISBN 9780197507629 (hardback) | ISBN 9780197507643 (epub) | ISBN 9780197507650
Subjects: LCSH: Antitrust law—Economic aspects—United States. | Labor economics—United States.
Classification: LCC KF1652 .P67 2021 | DDC 343.7307/21—dc23
LC record available at https://lccn.loc.gov/2021017327
DOI: 10.1093/oso/9780197507629.001.0001 1 3 5 7 9 8 6 4 2
Printed by LSC communications, United States of America
Introduction
In the United States, and much of the Western world, economic growth has slowed, inequality has risen, and wages have stagnated. Academic research has identified several possible causes, ranging from structural shifts in the economy to public policy failure. One possible cause that has received increasing attention from economists is labor market power, the ability of employers to set wages below workers’ marginal revenue product.1 New evidence suggests that many labor markets around the country are not competitive but instead exhibit considerable market power enjoyed by employers, who use their market power to suppress wages. This phenomenon—the power of employers to suppress wages below the competitive rate—is known among economists as labor monopsony, or simply labor market power. Wage suppression enhances income inequality because it creates a wedge between the incomes of people who work in concentrated and competitive labor markets. Wage suppression also reduces the incomes of workers relative to those of people who live off capital, and the latter are almost uniformly wealthier than the former. Wage suppression also interferes with economic growth since it results in underemployment of labor and, while it may seem to raise the return on capital, actually depresses it, as capital must lie idle to take advantage of monopsony power. With wages artificially suppressed, qualified workers decline to take jobs, and workers may underinvest in skills and schooling. Many workers exit the workforce and rely on government benefits, including disability benefits that have become a hidden welfare system.2 This in turn costs
the government both in lost taxes and in greater expenditures. One estimate finds that monopsony power in the U.S. economy reduces overall output and employment by 13% and labor’s share of national output by 22%.3
The claim that labor market power raises inequality and reduces growth mirrors another claim that has received attention lately—that the product market power of firms has contributed to rising inequality and faltering growth.4 A product market is a collection of products defined by frequent consumer substitution. When a small number of sellers or one seller of these products exist, we say that each seller has product market power, which enables it to charge a price higher than marginal cost, or the price that would prevail in a competitive market. When a small number of employers hire from a pool of workers of a certain skill level within the geographic area in which workers commute, the employers have labor market power
One major source of market power in both types of markets is thus concentration, where only a few firms operate in a given market. Imagine, for example, a small town with only a few gas stations. Each gas station sets the price of gas to compete with the prices of the other gas stations. When a gas station lowers its price, it may obtain greater market share from the other gas stations—which increases profits—but it also receives less revenue per sale. If only a single gas station exists, it will maximize profits by charging a high (“monopoly”) price because the gains from buyers willing to pay the price exceed the lost revenue from buyers who stay away. If only a few gas stations exist, they might illegally enter a cartel in which they charge an above-market price and divide the profits, or they might informally coordinate, which is generally not illegal, though the social harm is the same. In contrast, if many gas stations compete, prices will be bargained down to the efficient level—the marginal cost—resulting in low prices for consumers and high aggregate output of gasoline.
Labor market concentration creates monopsony (or, if more than one employer, oligopsony, but I use these terms interchangeably) where labor market power is exercised by the buyer rather than (as in the example of gas stations) the seller. Employers are buyers of labor who operate within a labor market. A labor market is a group of jobs (e.g., computer programmers, lawyers, or unskilled workers) within a
geographic area where the holders of those jobs could with relative ease switch among the jobs. The geographic area is usually defined by the commuting distance of workers. A labor market is concentrated if only one or a few employers hire from this pool of workers. For example, imagine the gas stations employ specialist maintenance workers who monitor the gas-pumping equipment. If only a few gas stations exist in that area, and no other firms (e.g., oil refineries) hire from this pool of workers, then the labor market is concentrated, and the employers have market power in the labor market. To minimize labor costs, the employers will hold wages down below what the workers would be paid in a competitive labor market—their marginal revenue product. Faced with these low wages, some people qualified to work will refuse to. But the employers gain more from wage savings than they lose in lost output because of the small workforce they employ.
Antitrust law does not distinguish monopoly and monopsony (including labor monopsony): firms that achieve monopolies or monopsonies through anticompetitive behavior violate antitrust law. But product market concentration has received a huge amount of attention by courts, researchers, and regulators, while labor market concentration has received hardly any attention at all.5 The Department of Justice (DOJ) and Federal Trade Commission’s (FTC) Horizontal Merger Guidelines, which are used to screen potential mergers for antitrust violations, provide an elaborate analytic framework for evaluating the product market effects of mergers. Yet, while the Merger Guidelines state that there is no distinction between seller and buyer power,6 they say nothing about the possible adverse labor market effects of mergers. Similarly, while there are thousands of reported cases involving allegations that firms have illegally cartelized product markets, there are few cases involving allegations of illegally cartelized labor markets.7
This historic imbalance between what I will call product market antitrust and labor market antitrust has no basis in economic theory. From an economic standpoint, the dangers to public welfare posed by product market power and labor market power are the same. As Adam Smith recognized, businesses gain in the same way by exploiting product market power and labor market power—enabling them to increase profits by raising prices (in the first case) or by lowering costs
(in the second case).8 For that reason, businesses have the same incentive to obtain product market power and labor market power. Hence the need—in both cases—for an antitrust regime to prevent businesses from obtaining product and labor market power except when there are offsetting social gains.
Why, then, the imbalance between product and labor market antitrust? There are a few possible answers. First, while economic theory treats product markets and labor markets similarly, legal theory has placed more emphasis on product markets. One possible reason for this is that since the 1960s, legal scholars have influentially argued that the amorphous norms of antitrust law that prevailed earlier in the 20th century should be replaced with a laser-like focus on consumer welfare.9 The resulting shift in focus naturally favored product market analysis because consumers primarily are injured by price increases caused by product market power. In contrast, primarily workers are injured by the exercise of labor market power. Of course, workers are consumers, and so workers benefited from the law’s attention to product markets—but not as much as they would have if the law had paid attention to labor markets as well.
Second, economists until recently assumed that labor markets were approximately competitive, and accordingly that labor market power was not an important social problem. Most people live in urban areas where numerous employers vie for workers. Workers can (and do) move around the country if jobs are scarce or pay is low where they live, putting an upper bound on the social cost of labor market power. A well-known textbook on industrial organization, published in 2005, claimed, “Most labor economists believe there are few monopsonized labor markets in the United States.”10 It is only in recent years that these assumptions have been thrown into doubt. Moreover, academic economics has long been divided into separate fields of industrial organization (IO) and labor economics. IO economists have focused on antitrust problems created by mergers and other corporate actions, while labor economists have focused on labor and employment law. Partly as a result, labor economists never developed the analytic tools relevant to forecasting the impact of increased labor market power that are analogous to or draw on the models IO economists use to analyze product market power.
The DOJ and FTC rely heavily on advice from economists on their staff when evaluating mergers and have frequently challenged mergers based on their effects on product markets. Likely relying on the traditional assumption of economists that labor markets are competitive, the agencies have never blocked a merger because of its effect on labor markets.
Third, antitrust litigation against employers is more difficult than antitrust litigation based on product-market concentration, perhaps giving the illusion that the latter problem is more significant than the former. Product-market litigation is often brought by large firms, which have the resources and incentives to bear the high costs of complex and expensive antitrust litigation. Class actions by consumers are also relatively straightforward because in a typical antitrust case involving product markets, the argument is simply that the consumers paid a higher price than they should have, which means that the consumers share a common interest as required by courts.11 In contrast, virtually no worker can hope to obtain damages in an antitrust action—even with the treble damages rule—that would compensate her for the cost of litigation. And class actions brought by workers face significant obstacles because workers—unlike consumers—are frequently in diverse positions, defeating the common interest requirement for class certification.12
That was the intellectual and legal landscape until a few years ago. The consensus that labor markets were competitive collapsed in response to several events. First, there was the revelation that high-profile Silicon Valley tech firms, including Apple and Google, entered nopoaching agreements, in which they agreed not to solicit each other’s employees.13 This type of horizontal agreement is a clear violation of the antitrust laws. The firms settled with the government, but the casual way in which such major firms, with sophisticated legal staffs, engaged in such a blatant violation of the law alarmed antitrust authorities. The government subsequently issued guidelines to human resources offices warning them that even implicit agreements not to poach competitors’ employees were illegal.14 In 2016 the White House Council of Economic Advisors and the Department of the Treasury issued reports warning of the dangers of cartelized labor markets.15 The DOJ has also launched criminal investigations of firms for entering no-poaching agreements.
Shortly before this book went to press in 2021, the DOJ announced its first criminal indictments of participants in a pair of wage-fixing and no-poaching conspiracies.
Second, the recent discovery that noncompete agreements—which prevent workers from moving from their employer to a competitor— are extraordinarily common and frequently are applied to low-wage workers raised suspicions that they were being used by employers to exploit their labor market power. For example, Jimmy John’s, a sandwich franchise, routinely required low-wage employees to sign covenants not to compete, which apparently deterred those employees from moving to competitors.16 While these noncompetes were probably unenforceable under the common law, the workers lacked the wherewithal to consult lawyers and threaten (or defend against) lawsuits. The effect was to reduce mobility between jobs, possibly suppressing wages. Researchers subsequently learned that an enormous number of workers—including low-income, relatively unskilled workers—are bound by restrictive covenants. According to one study, in 2014 12% of workers earning less than $40,000 per year with education below the college level were bound by noncompetes.17 Relatedly, large franchises like McDonald’s used no-poaching agreements to reduce competition among franchisees for workers—which in some markets might have resulted in considerable increases in market power.18 News media reports provide additional anecdotal evidence of the ubiquity of noncompetes and no-poaching agreements and their powerful effect on labor mobility.19
Third, economists began investigating monopsony in labor markets. An important spark for this work was a classic study by David Card and Alan Krueger, which found that employment levels were not affected by a minimum wage hike in New Jersey in 1992.20 While controversial at the time, many other studies of minimum wage increases in other jurisdictions and at other times produced similar results.21 A possible explanation for the result is that labor markets are monopsonies: if employers pay workers less than their marginal product, then a minimum wage hike—if not too great—will result in higher wages without disemployment.
While other explanations are possible, the monopsony theory gains credence from other studies of the past several years, in which economists, using a range of methodologies as well as previously
unavailable sources of data, have found evidence of widespread labor market concentration. I will discuss these studies in chapter 1. The key point for present purposes is that the studies show that most labor markets are highly concentrated and suggest that labor market concentration results in lower wages, as one would expect. In recent years, with the focus on Google, Facebook, and Amazon, people have begun to believe that market concentration is a problem for American consumers. But if market concentration is a problem, it is a bigger problem for workers than for consumers.
This book argues that antitrust law should be brought to bear against labor monopsony. I argue that the historical neglect of labor monopsony by antitrust law is not justified by theory, law, or the nature of labor markets. Antitrust law should police labor markets just as it polices product markets. Indeed, the case for antitrust law enforcement in labor markets is stronger than in product markets. Labor markets are likely to be more concentrated than product markets and more vulnerable to anticompetitive behavior by firms. To this end, I lay out the ways in which antitrust law can be adapted to the particularities of labor markets. While I propose some reforms, my main argument is that antitrust law as it is understood today has been underused in labor markets—that good lawyers and sophisticated regulators and judges who are attentive to the special features of labor markets could make progress even without legal reform.
A few words about this book’s methodology. Antitrust law is currently embroiled in controversy about its assumptions, goals, and methods. This book does not take sides in this debate. My argument uses the traditional antitrust methods that are currently dominant in courts and academic scholarship. For workers, the problem has not been that antitrust law has the wrong economic goals; it is that antitrust has almost never been applied to labor markets.
And now a word on vocabulary. In antitrust law and IO economics, a firm is a monopoly if it is technically the only seller in a market and, more familiarly, the dominant firm in a market that has multiple firms. While antitrust lawyers and IO economists use the word “monopsony”
in a similar way to refer to a single buyer from a product market, in labor economics the term is confusingly used in a more general way to refer to a market in which employers (buyers of labor) pay a wage below the competitive level (the worker’s marginal revenue product). Monopsony can be the result of search frictions and other factors. The two groups of economists use the term “market power” in the same way. Product market power means the power to raise prices above the competitive rate; labor market power means the power to pay wages below the competitive rate. Thus, monopsony is a synonym for labor market power, while monopoly is just one source of product market power rather than a synonym for it. I will use the word “monopsony” in both ways: to refer both to a buyer (employer) with labor market power (whatever the source) and to a buyer (employer) who dominates the labor market as a sole or very large employer. Context should keep things clear.
Labor Monopsony in the United States
1.1. The Intellectual History of Monopsony
The term “monopsony” was coined by the British economist Joan Robinson in 1933.1 Before then, economists who wrote about industry structure focused on monopoly and market power in the product market. It had long been clear that large corporations like Standard Oil monopolized goods and services. Robinson realized that corporations also exercised market power on the “buy side”—in their purchases of inputs, including goods, services, and labor. The monopsony power of corporations was just as common as their monopoly power, perhaps more common, but harder to detect.
To be sure, long before Robinson’s book, economists and other thinkers understood that employers exercised power over workers. Adam Smith observed that “masters” (employers) tacitly “combined” or agreed not to raise the wages of workers. Anticipating the silence about labor monopsony even today, he wrote:
We seldom, indeed, hear of this combination, because it is the usual, and one may say, the natural state of things which nobody ever hears of. Masters too sometimes enter into particular combinations to sink the wages of labour even below this rate. These are always conducted with the utmost silence and secrecy, till the moment of execution, and when the workmen yield, as they sometimes do, without resistance, though severely felt by them, they are never heard of by other people.2
Later, critics of capitalism like Karl Marx denounced employers’ treatment of their workers. They argued that employers exploited their workers by underpaying them, subjecting them to substandard working conditions, and busting unions. Marx argued that employers could keep wages low by taking advantage of what he called the “reserve army of the unemployed.”3 Because people desperately sought work, employers could keep workers’ wages below the value of their labor for the employer or for society, at their “labor power” (the minimum they needed to continue working). The extraction of the resulting “surplus value” by the employer was what Marx called “exploitation.” Robinson’s analysis of monopsonistic labor markets provided a more rigorous formulation of the problem of employer dominance. She pointed out that if labor markets are competitive, employers cannot exploit workers in Marx’s sense because workers who are paid only enough to avoid starvation will be able to sell their labor to other employers at a higher wage. Yet labor markets need not be competitive, and when they are not, the outcome for workers is similar to those that Marx identified: excess unemployment, a permanent gap between wages and worker productivity, poor working conditions, and domination of the workers by employers.
Robinson’s terminology was adopted by some scholars in the institutionalist tradition in labor economics. For example, in 1946 Lloyd Reynolds, an early founder of the subfield of labor economics, published two noteworthy papers, “The Supply of Labor to the Firm” and “Wage Differences in Local Labor Markets.”4 Both of these papers described empirical features of the labor market (such as wage differences for similar workers within a local labor market) that were consistent with employer market power. Reynolds anticipated virtually all of the modern economic models of monopsony.
While Marx held that the increasing concentration of markets under capitalism would spark a revolutionary transition to communism, and other thinkers hoped that workers would be able to govern themselves in communes and other organizations, the major effect was instead unionization, which, while often militant, was focused on improving wages and conditions for workers rather than overthrowing governments or social orders. In the United States and other Western countries, governments initially resisted the labor movement, often
violently, but ultimately accommodated it. Governments passed laws that protected workers from low wages, excessive hours, dangerous workplaces,5 and other abuses, and that protected labor organization from interference by employers.6
The bottom-up nature of worker organization probably accounts for a major bifurcation in the way American policy and law treat the problem of labor market power. While antitrust law, beginning with the Sherman Act of 1890,7 nominally applied to all markets—labor markets as well as product markets—antitrust enforcement would focus almost exclusively on product markets.
Meanwhile, a separate body of law emerged for addressing labor markets. Unions were initially ambivalent toward antitrust law. Union leaders worried that if they made demands on weak employers in a competitive market, the employers would go bankrupt. Better to negotiate with the industry as a whole, and even a single firm that dominated the industry, as long as unions could represent the entire workforce. This ambivalence toward antitrust law turned to hostility when courts initially held that unions themselves could violate the antitrust laws since they combined to fix the price of labor. In 1914, Congress passed the Clayton Act, which exempted unions from antitrust law, but suspicion toward antitrust law lingered. Unions sought and finally obtained—with the National Labor Relations Act of 1935—formal legal recognition by the government and the right to strike and to bargain collectively. Labor organization offered an alternative to antitrust law: rather than break apart employers into competitive buyers of labor, unions bring together workers so that their aggregated bargaining power could counter the bargaining power of the large employer.
This bifurcation of the law led to a bifurcation in economic theory. To address monopolization of product markets, economists developed the field of industrial organization (IO), which seeks to explain how market power affects the structure of business mainly in relation to product markets. To address abuses in the labor market, economists developed the field of labor economics, which focuses on unions and employment regulations. Product markets and labor markets differ in many ways, and economists focused on the most salient aspects of them. IO economists focused on “market structure”: why goods and services were supplied in different types of markets, which firms produce the
goods and services, and so on. Labor economists focused on training, job mobility, unionization, and related phenomena.
In the United States, the power of unions peaked in the 1950s and has steadily declined ever since. Foreign competition and technological change played a role in the decline of unions. So did the advance of union-busting tactics and rising hostility from business, government, and public opinion. Beginning in the 1970s and accelerating in the 1980s and 1990s, public policy turned against market regulation—both antitrust law and product market regulation, and labor law and labor market regulation. The “neoliberal” revolution reflected frustration with the rigid, outmoded character of traditional market regulation. Yet rather than adapt new laws or organizations better suited to the age of globalization and digitization, reformers focused on dismantling existing market regulation.
1.2. The Sources of Monopsony Power
Economists have identified several sources of labor market power. Because these sources have counterparts in the more familiar analysis of product markets, I will introduce each of them by way of the product market.
In product markets, there are three primary barriers to competition. First, market concentration refers to the existence of one or a small number of sellers, usually the result of increasing returns to scale, high fixed costs, or network effects. Second, product differentiation exists when goods or services are different from each other rather than fungible; differences across products make comparison difficult, which reduces competition. Third, search frictions make it hard for consumers to compare products and seek out the best offering. In both the academic literature and legal adjudication, market concentration typically plays the central role in analysis, with less emphasis on product differentiation and the least emphasis on search. In the literature on labor markets, by contrast, the order of emphasis has been the reverse.
Search frictions. Search frictions or costs refer to the difficulty of finding a job. We all know from experience that finding a new job is not like shopping for clothes or furniture; job hunting requires considerable effort, including research, travel, and interviewing. The problem
of search frictions has played the central role in understanding labor markets.8 In 1998, Kenneth Burdett and Dale Mortensen proposed a model of labor markets with a large number of identical workers and identical firms where search frictions naturally lead employers to have monopsony power.9 Alan Manning, in his 2003 book Monopsony in Motion, presented a wide variety of evidence in favor of what is called the dynamic monopsony model.10 Because a worker’s existing employer knows that the worker’s search cost is high, the employer can reduce compensation—including wages, benefits, and workplace amenities— or fail to increase compensation despite the worker’s contributions because the employer knows that the worker can find an alternative job only with difficulty.
Job differentiation. In recent years, labor economists have focused on firm-specific amenities of a workplace, which is the labor market correlate to product differentiation.11 Imagine two workplaces that are identical at an initial point. The employer of each workplace seeks to deter workers from leaving. To do so, the employer might offer an amenity that its workers happen to like—say, a coffee bar or a yoga studio or hot showers. While these amenities may seem frivolous, many of the most important amenities are extremely significant but less apparent as they are “omissions” rather than “commissions”; many jobs involve odd hours and unpleasant or hazardous working conditions. The absence of such “dis-amenities” itself makes jobs attractive. Other amenities might rise more naturally; for example, the location of an employer might appeal to workers because of the convenience to commuting or the attraction of nearby restaurants or other businesses. Differing amenities give rise to search frictions, as noted above, but they separately make it more difficult for workers to compare firms. Indeed, the identities of the other workers at a workplace—whether they are driven and intense, friendly and laid back, or young or old—matter to people, and even very similar-seeming employers, for example, law firms, might be very different in practice. Other common amenities include shift flexibility, childcare, vacation and sick time, and opportunities for promotion and personal growth.
Not only are jobs differentiated along these various dimensions, but workers perceive and value these dimensions differently. Even jobs that seem quite standardized (e.g., cashier) in terms of the productive tasks
are frequently valued idiosyncratically by workers who have particular tastes over commutes and workplace relationships. Crucially, firms may not be able to observe this taste heterogeneity. This limitation, plus practical and moral constraints that require firms to post only one wage per job, is what makes labor market power inefficient: if firms could perfectly tailor the wage to each worker’s taste for working at that firm, so-called perfect wage discrimination, there could still be market power, but it would not be inefficient, even as all surplus is extracted by the employer.
A major but now forgotten development in labor relations in the United States was the bureaucratization of labor relations, which took place roughly from World War I though the Great Depression.12 Often at the instigation of the government, which had considerable authority over employers during World War I, employers agreed to regularize work positions. Job categories were defined for the first time; each job category involved specified skills and specified tasks, and workers within each job category received the same pay, adjusted for seniority. Work would take place at certain hours and in certain locations. Job categorization improved efficiency because the publicly stated criteria made it easier to compare and evaluate workers and to transfer them from job to job. Unions supported job categorization because it facilitated bargaining at the level of the bargaining unit; employers would have more difficulty undermining unions by subtly adjusting their pay and conditions of work for workers, creating divisions within the union. Employers were less enthusiastic about job regularization but gave in to pressure from the government and unions and appeared to see some of the benefits of the bureaucratized system. But a key element of job bureaucratization about which employers had no enthusiasm is that it made it easier for employers to compete for workers at other firms since their job category would render workers interchangeable with similar work positions at the poaching firm. The reduction of job differentiation would thus have weakened the employer’s labor monopsony.
Employers may have sought to strengthen their labor monopsonies by advancing what was then known as “welfare work,” a style of management that took a paternalistic interest in the well-being of workers and their families.13 The employer sought to create a sense of “family” in the workplace by emphasizing personal ties in the workplace and boosting
team spirit with “company picnics, company athletics, company songs, company contests, and company magazines filled with inconsequential gossip about these activities.14 The economists Samuel Bowles and Herbert Gintis would later emphasize the role of endogenous worker preferences in maintaining employer profits.15 A large literature in sociology, perhaps beginning with Michael Burawoy’s 1979 Manufacturing Consent: Changes in the Labor Process under Monopoly Capitalism, argues that employers engineer work environments so as to maximize worker productivity and minimize conflict with management.16 Workers also develop attachments to coworkers, employers, customers—even the job itself.17 Employers may try to invest in increasing these attachments by fostering corporate loyalty or specializing the content of training programs.
Labor market concentration. The final and most neglected cause of labor market power is the concentration of labor markets as a result of economies of scale, network effects, fixed costs, and other factors. In many industries a firm with many employees can churn out goods and services more efficiently—at less cost per unit of output—than firms with fewer employees can. Firms can grow large naturally—as they take on additional business and hire workers to supply labor for production or services—or by merging. When a single firm or a small number of firms hire from a labor market, those firms have labor market power. A single firm maximizes profits by choosing a wage below the competitive wage. Since workers cannot quit and find a competing employer to hire them, they must either accept the wage or undergo the costly process of dropping out of the labor market and either retiring or retraining.
When more than a single employer hires from a labor market, but the number of employers remains small (or a few of the employers are dominant while others are minor), employers maximize their profits by paying a wage somewhere between the competitive and monopsony levels. This phenomenon is familiar from the product market. A small number of sellers can engage in parallel pricing or tacit collusion in order to keep prices high. Rather than set prices at marginal costs, they set prices to match those of a price leader (typically the biggest firm) or simply match the prices of their competitors. Employers do the same thing. Large or otherwise salient employers announce wages, and their
competitors follow them. Or employers keep wages constant (or adjust them mechanically for cost of living) and raise them only when workers can credibly claim that another employer will pay them more. As long as the number of employers is not too large, they can maintain wages below the competitive rate.18
Table 1.1 summarizes the three main sources of labor market power.19
While the sources of labor market power are different, their effect is the same.20 The employer pays a wage below the competitive rate, and employers hire workers at below the competitive quantity. This results in two negative effects for public policy: a loss of output (or efficiency) and a harm to equity if (as is almost always the case) workers are less wealthy than shareholders.
The discussion so far might give the impression that labor markets and product markets are similar: they are both vulnerable to market power (monopsony in the first case, monopoly in the second), and for the same reasons. But there is reason to believe that labor markets are more vulnerable to monopsony than products markets are to monopoly. Labor markets, much more than product markets, are characterized by matching. 21 The preferences of both sides of the market affect whether a transaction is desirable.
Compare buying a car in the product market and searching for a job. Both are important, high-stakes choices that are taken with care. However, there is a crucial difference. In a car sale, only the buyer cares about the identity, nature, and features of the product in question: the car. The seller cares nothing about the buyer or what the buyer plans to
Sources of monopsony power
Search frictions
Job differentiation
Concentration
Meanings
Workers have difficulty learning about comparable jobs.
Similar-seeming jobs are actually different because amenities are different, including conditions, commute.
Few employers offer a certain kind of job.
TABLE 1.1 Sources of Monopsony Power
do with the car. In employment, the employer cares about the identity and characteristics of the employee, and the employee cares about the identity and characteristics of the employer. Complexity runs in both directions rather than in one. Employers search for employees who are not just qualified but who possess skills and personalities that are a good match with the culture and needs of that employer. At the same time, workers look for an employer with a workplace and working conditions that are a good match for their needs, preferences, and family situation. Only when these two sets of preferences and requirements match will a hire be made. This dual set of relevant preferences means that most labor markets are doubly differentiated by both the idiosyncratic preferences of employers and those of workers.
These matching frictions both cause and reinforce the typically long-term nature of employment relationships compared to most product purchases, leading to significant lock-in within employment relationships. They are also reinforced by the more geographically constrained nature of labor markets. Products are easily shipped around the country and the world; people are not. While traveling is easier than in the past, and remote work has become more common in the wake of the COVID-19 pandemic, labor markets remain extremely local, while most product markets are regional, national, or even global. Most jobs still require physical proximity to the employer, greatly narrowing the geographic scope of most labor markets, given that many workers are not willing to move away from family to take a job. Two-income families further complicate these issues because each spouse must find a job in the geographic area in which the other can, further narrowing labor markets. Together these factors naturally make workers highly vulnerable to monopsony power, much more vulnerable than most consumers are to monopoly power.
1.3. Measuring Market Power
The economics of labor market power and the harms it causes are closely analogous to the theory of product market monopoly. A monopolist is a firm that does not take market prices as given, but can raise its price, at the cost of some lost sales, to increase the profits it earns on each sale. In choosing an optimal price the firm faces a trade-off.
Raising the price reduces sales but also increases the revenues the firm earns on each unit sold. The higher the firm raises its price, the costlier it is to lose extra demand, as each sale is very profitable. Eventually the firm finds a balance point, where the value of the lost sales from raising the price further just equals the increased profits on the units it sells at the increased price. The firm’s absolute markup is the gap between this price and the firm’s cost. The markup equals the difference between the monopoly price and the competitive price, and thus serves as a natural gauge of market power.
A similar trade-off between profit-per-unit and number of units sold applies to firms that are not literally monopolists but have some power over their price. In fact, in some sense, every firm with any market power is a “monopolist” over some market, though maybe one too narrow to matter much.
The analysis of monopsony in labor markets is analogous. In a competitive labor market, firms equate the going wage of workers to their “marginal revenue product,” the amount of additional revenue the worker can generate. When an employer has a monopsony, it considers the fact that to hire an additional worker it will have to raise the prevailing level of wages for its existing workers and that doing this will increase its overall labor costs. The reason (as noted above) is that employers typically cannot wage-discriminate: they must pay the same wage for the same job. They cannot wage-discriminate because they cannot determine the worker’s reservation wage (the wage below which she will quit) and because workers get very angry when they learn that colleagues with the same skill and experience are paid more than they are (known as the pay equity norm).22
Conversely, if an employer lowers wages, while it will lose some workers, it will also lower the wage bill on the workers it already employs. As in the monopoly case, a monopsonist will not internalize this effect on workers and will choose a markdown of wages below the marginal revenue product. Just as with firms with product market power, an employer with labor market power may not have a “monopsony” over some easily described market, but so long as it will not lose its entire workforce by slightly lowering its wages, it has some labor market power.23
Economists use the word “elasticity” to refer to the sensitivity with which one economic variable changes in response to a change in another. “Labor supply elasticity” refers to the sensitivity with which workers react to changes in wages. Suppose that wages across the economy decline a tiny amount, and everyone quits. Then labor supply elasticity is (at the limit) infinity or, in other words, very high. Suppose instead that no one quits; then labor market elasticity is zero. Elasticity can range from zero to infinity. An elasticity of 1 means that if wages increase, say, 1%, then employment will increase 1% as well. An elasticity of 0.5 means that a 1% wage increase will result in a 0.5% employment gain, while an elasticity of 2 means that the same wage increase will result in a 2% employment gain. A relatively high level of elasticity indicating a competitive market for practical or policy purposes might be around 10; the 1% wage increase results in a 10% employment gain. The general view is that labor supply elasticity across the economy is in the neighborhood of 0.5, suggesting a high level of inelasticity.24 People tend to stay in the workforce even when wages decline below the competitive rate because they need to support themselves.
The term “residual labor supply elasticity” refers to the sensitivity with which workers react to changes in wages at a particular firm. Suppose a computer programmer who works at Google would quit and move to Apple if wages at Google decline by a tiny amount. Then the residual labor supply elasticity is high. If the programmer would not quit even if Google lowered wages significantly, then the firm-level elasticity is closer to zero. Like labor supply elasticity (sometimes called “aggregate labor supply elasticity” to distinguish it from “residual labor supply elasticity”), residual labor supply elasticity can fall anywhere along this continuum, though it can never fall below the (aggregate) labor supply elasticity for the relevant category of workers. But it varies greatly from industry to industry.25
Residual labor supply elasticity is a simple measure of a firm’s labor market power. If workers do not quit even if the firm lowers wages significantly (elasticity is low), then the firm enjoys significant market power over the workers. This is the number that antitrust policy focuses on. If residual labor supply elasticity that a firm faces is high, then the
labor market from which a firm draws its workers is competitive, and the firm cannot suppress wages below the competitive wage, or not by much. If it is low, workers need protection.
The economic consequences of labor market power are analogous to those of product market power. Product market power has two wellknown effects. It redistributes from consumers to the firm: consumers must pay more for products, and the firm earns greater profits at their expense. And it creates waste or deadweight loss. Some consumers would be willing to pay the efficient, marginal cost price that the firm would have charged in a competitive market but are not willing to pay the higher price the monopolist chooses to charge.
Similarly, monopsony power has two effects. It redistributes from workers to employers by lowering wages. And it creates waste: some workers would have been willing to work for the employer if they had been paid their full marginal revenue product but will quit if they are paid the marked-down wage the monopsonist offers. This leads to increased unemployment or nonemployment as workers find prevailing wages unacceptable and exit the labor force or refuse to take available jobs. Economic output also declines.
Monopsony power creates other negative effects as well. First, to the extent that the degree of monopsony power differs across employers, it will also lead to misemployment: workers may be more productive at employer A, which has a lot of labor market power, than at employer B, which has a little. But B may offer higher wages because of its limited labor market power. The worker may thus choose to work at B, lowering the productivity of the economy. Misallocation may be particularly severe because of the two-sided matching problem. If matches between workers and firms generate specific benefits, monopsony can distort which firms match which workers, which will lower the allocative efficiency of the market.
Second, employers will often cut benefits, rather than cut wages, to take advantage of workers who are locked into the job. The firm has no need to retain these workers and thus may wastefully degrade conditions of work these “stuck” workers particularly value, instead catering only to the workers the firm is worried about losing.26
Third, monopsony raises prices for consumers. This may seem counterintuitive: won’t lower wages to workers be passed through to
consumers as reduced prices? That argument is often made as a defense of monopsony power.
In fact, however, this argument is wrong. To see this, note that if firms employ fewer workers, they will produce less output, resulting in higher prices. The labor cost savings accrue to the employer itself (or its shareholders), not to the buyers of its goods. Those buyers will pay a price that is determined by the structure of the product market, not the labor market. So, for example, if the employer is also a monopolist in the product market, it will charge the buyers the monopoly price—which is determined by how much buyers are willing to pay. And if the product market is competitive, the employer will charge prices for its goods that are no higher than the competitive price—with its competitors taking up the slack as the employer itself will produce less given its small workforce. The technical explanation is that while the firm lowers wages to workers, the cost to the firm of hiring workers rises as the firm now considers the fact that, when it hires an additional worker, it also will pay its other workers more. When a monopsonist hires a single worker, it must increase wages for all its workers. (Recall that employers cannot easily wage-discriminate.)27 If this seems paradoxical, note that it is merely the flip side of a well-understood feature of monopolistic control of product markets: that a monopolist produces fewer products and charges a higher price for them than does a competitive firm. Monopoly and monopsony are two sides of the same coin, and both harm labor and product markets.
Fourth, and precisely for this reason, monopsony power reinforces and exacerbates monopoly power. In fact, both can be seen as two alternative ways for the owners of capital to squeeze workers and thus reduce the returns to productive work and the output of the economy. The markdown on wages caused by monopsony and the markup on prices caused by monopoly are akin to taxes: payments that ordinary people must pay in order to go about their daily life as producers and consumers. However, the payments go not to governments to fund programs, but to firms and, ultimately, investors. And the payments do not spur investment and raise economic growth because they depend in the first place on the willingness of managers to leave capital idle to obtain market power, while driving workers out of the workforce and onto taxpayer-financed relief programs.
1.4. Evidence of Labor Market Power
Until recently, economists entertained the possibility that labor monopsony could exist, but data limitations prevented any consensus from emerging about its prevalence.28 The field seemed to conclude that while search frictions could give labor market power to employers, concentration and collusion were not a significant source of labor market power. Historical studies of company towns found weak evidence that they conferred labor market power to employers.29 Studies of labor markets for nurses found stronger evidence of anticompetitive labor market practices but were vulnerable to methodological objections.30 There was little doubt that Major League Baseball and other sports leagues engaged in anticompetitive labor market practices, but the unusual nature of these organizations suggested that labor market collusion was anomalous.
Yet there are powerful indications that employers have colluded in labor markets throughout history, often with the help of governments. In England, anticompetitive elements of master and servant law dated back to the 1351 Statute of Labourers.31 These rules included enticement doctrines, which blocked employers from poaching each other’s workers, and criminalization of worker “absconding,” i.e., leaving an employer without permission. The feudal legacies may have legitimized anticompetitive practices in labor markets.32 In the United States, of course, literal ownership of human beings was common in the South until the Civil War put an end to slavery. In the wake of the war and ratification of the 13th Amendment, courts were skittish about enforcing restraints of trade in the labor market. But in the postbellum U.S. South Jim Crow labor law reintroduced anti-enticement rules and in other ways restricted the labor freedom of Blacks.33 In the North, the labor market power of the trusts was a common source of concern.34 Nineteenthcentury labor activists identified market power of employers as a justification for collective bargaining and regulation, complaining of “wage slavery” and drawing analogies to chattel slavery.35
The best data, and hence the best studies, come from the past few decades, and especially the past few years. The evidence consists of studies of many different markets, which tend to show that residual labor market elasticities are extremely low. One should acknowledge
