Economist’s perspective: US antitrust regulators focus on defining relevant labour markets
This is an Insight article, written by a selected partner as part of GCR's co-published content. Read more on Insight
This article describes economic issues that arise in defining antitrust labour markets and how courts have evaluated these in recent and ongoing cases concerning low-skilled workers.
- Increased focus of US regulators on antitrust issues in labour markets
- Economics of labour market monopsony
- Market definition in labour markets
- Court rulings on relevant labour markets for lower-skilled workers in two recent cases
Referenced in this article
- Deslandes v McDonald’s USA et al., No. 17-CV-4857 ND Ill, Eastern Division
- Jien v Perdue Farms, Inc et al., No. 1:19-CV-2521-SAG, D Md
- US Department of the Treasury, ‘The State of Labor Market Competition’
- Department of Justice Antitrust Division and Federal Trade Commission, ‘Antitrust Guidance for Human Resource Professionals’
US antitrust regulators and policymakers have identified protecting competition in labour markets as a priority. With the increased focus on competition in labour markets, attention is focused on relevant labour market definition to assess whether potentially anticompetitive conduct would have appreciable effects. In recent antitrust actions involving fast food restaurants and poultry processing, courts have considered market definition as an important step to assessing market power. These cases differ from previous high-profile actions in that the jobs in question generally do not require specialised training or experience.
A relevant labour market is one in which, from the perspective of workers, jobs are reasonably interchangeable with the jobs at issue in the matter. In broad terms, the relevant jobs are those that workers are willing and able to do in terms of skills and commuting distance; in other words, the relevant labour market has a product or service and geographic dimension. Just as consumers’ willingness to substitute among products in the relevant market disciplines firms and prevents them from raising prices, in principle, the jobs offered by competing employers in a relevant labour market are those that workers view as reasonably interchangeable, and therefore workers’ willingness to move between jobs will constrain firms’ wage-setting decisions in the labour market.
This article examines these issues in the context of recent experience with market definition from cases that involve demand for relatively low-skilled workers. We describe the economic issues and the recent cases where market definition has played a significant role. Based on our review, we observe:
- as with product market definition, substitution – in this case, job switching by workers – has a central role in labour market definition;
- labour markets for lower-skilled workers will often include more substitutes than the corresponding product (or output) markets because of the nature of labour market relationships and the multiple dimensions on which workers evaluate alternatives; and
- the nature of labour markets can make it harder to measure substitution patterns and other features. In many markets, workers tend to switch jobs infrequently, have limited information on available choices, and any change in jobs requires two-sided matching, which means that employer decisions will also be important for labour market definition.
Market definition and market power
Defining the relevant market is an initial step in the structural approach to assessing market power because, without defining a relevant economic market it is not possible to compare the relative sizes of suppliers or buyers and thus start to consider the potential dependence of third parties on these suppliers or buyers. Market power can be evaluated indirectly via market structure, by determining whether the conditions and circumstances of the relevant market result in market power by the firm being scrutinised, or it can be evaluated directly, by identifying direct evidence of a firm’s ability to exercise labour market power in the form of lower wages or lower employment. Structural analyses, as the name implies, focus on measures such as market shares and concentration (ie, is the market in question dominated by one or a small number of suppliers or buyers?). A sound assessment of the relevant market is critical to the implementation of structural analyses because the share of the relevant labour market accounted for by the firm in the antitrust case are key inputs to the assessment of market power.
Under a structural approach, the existence or potential for a firm to exercise market power is inferred from the market shares of all relevant parties in the relevant market. Firms with larger shares are considered to be more likely to exercise market power. As a matter of economics, the relationship between market structure and market power is not straightforward for a number of reasons. First, firms may achieve a high market share without attaining substantial market power. For example, a first mover might achieve a high market share and still face pricing discipline from rival products if consumers view products as highly substitutable. Second, even within a relevant market, where firms are differentiated, some firms may be closer competitors than others (for example, in terms of their geographic proximity) in ways not captured by their overall market shares. Third, and on other hand, firms may face some competitive constraints from firms outside the relevant market. Moreover, inferring labour market power from market structure typically requires a detailed understanding of worker information and choices, employment frictions, employer needs and hiring criteria, and the dynamics of the economies in which workers are searching.
Labour market power
Labour market power, or monopsony power, is the ability to reduce wages relative to the level that would prevail in a competitive market. The concept of monopsony in a labour market is often presented in terms of a single employer, which can be thought of as the buyer-side analogy to the monopoly model with a single seller in an output market. In this representation, the single employer faces an upward-sloping labour supply curve because, as the only employer, to attract more labour it must increase the wage not just for the last worker hired but for all employed workers. Conversely, the monopsonist recognises that if it hires fewer workers, it can pay a lower wage for all the workers it employs. As a result, and analogous to the monopoly outcome, the single employer hires fewer workers than would be hired under perfect competition, and it pays a lower equilibrium wage. Output is reduced but the firm’s profit is increased because of the lower cost of production for all units sold.
More generally, the terms ‘monopsony power’ and ‘labour market power’ are used refer to any labour market departures from perfectly competitive outcomes. Most labour markets differ from the perfectly competitive benchmark in important ways. Labour markets involve two-sided matching: for a job to be filled, the employer must offer the position to a worker and the worker must accept. There are also significant frictions in labour markets: there are costs to employers of finding, hiring and training workers, and there are costs to workers of switching jobs (eg, the cost of finding a job that leaves them better off than the job they already have, including along non-monetary dimensions). Labour relationships typically are not purely transactional: employees return to work for the same employer over longer periods of time. These features alone can confer market power on employers and result in wages being below those that would prevail under the perfectly competitive labour market benchmark. However, this cannot be assumed as, by the same measure, employers may value reliable and high-effort workers. Efficiency wage theory suggests that where worker quality or effort is imperfectly observable, firms may pay a premium to secure and retain higher quality or higher effort workers.
Further complexities can arise if there are trade unions and collective bargaining by these unions because then even a monopsonist buyer may face a monopoly or quasi-monopoly supplier (depending on the degree of unionisation). In this scenario of ‘bilateral monopoly’ on both the buyer and supplier sides of the market, the precise outcome in terms of both wages and employment may depend on the relative bargaining strength of the firms and unions in question and their skill and sophistication in negotiations.
In practice, labour market definition involves a fact-intensive understanding of the jobs, their attributes and worker switching, as well as the geographic scope of markets. The fundamental aim of defining the relevant market is to identify the reasonably interchangeable alternatives available to workers. ‘From an antitrust perspective, firms that compete to hire or retain employees are competitors in the employment marketplace, regardless of whether the firms make the same products or compete to provide the same services.’ Information on reasonable substitutability of jobs can come from internal documents that show which jobs the company views as relevant for its employees and evidence that employers had offered raises to retain employees who had been offered positions at other employers. Data on job flows can also be used to identify worker alternatives. The US Census Bureau’s Longitudinal-Employer Household Dynamics data details worker-level information on employment, earnings and other features that can be used to measure job-to-job worker flows.
A set of substitute jobs, particularly for low-skilled workers, is often broader than a set of substitutes for the outputs they produce. In output markets, the majority of consumers typically choose between a small number of products that satisfy similar functions; however, the variety of jobs that employees come from or leave for is often much greater because skills are often transferable across industries. Moreover, when considering their options, workers may value job characteristics differently, affecting which jobs are closer substitutes (eg, some workers may value a predictable schedule highly while others may place premium on a short commute).
Research using the US Census Current Population Survey finds that just over 35 per cent of job switchers move to a job in the same four-digit industry while just under 35 per cent of switchers stay within the same three-digit occupation classifications. The corollary of these statistics is that at both the three and four-digit levels, some 65 per cent of job switchers find jobs in different industries, suggesting that neither level is a good guide to labour market definition. Therefore, industry codes and other pre-defined ways of classifying employment sectors should be used with care. On the one hand, they may be overly broad, and include employees and jobs that are not reasonably substitutable, and on the other hand, they may be overly narrow, and exclude employees and jobs that, to employees and employers, appear reasonably substitutable.
Other idiosyncrasies of labour markets should also be considered when ascertaining substitutes for specific jobs. Workers tend to switch employers relatively infrequently and, when they do, their choice set may be limited; whereas in output markets, consumers can buy whichever product they can afford. Job markets require matching a worker to an employer with an open position who is willing to hire that worker. As a result, the alternatives available to a worker (ie, the potential substitutes to a job) can change over time, in particular as the economy enters a different phase of the business cycle. Moreover, workers’ willingness to travel for new employment can lead to a relevant geographic market that differs from that of the output. For example, hospital executives may be willing to relocate across the country, suggesting a national geographic market, whereas the hospitals themselves compete for patients more locally. Conversely, low-skilled and less well-paid workers may not be willing to relocate to take advantage of regional wage discrepancies that are smaller in financial terms, whereas the output they produce is traded nationally. Finally, labour markets are characterised by information asymmetries: workers may be uncertain about working conditions at a prospective employer, and employers might not be able to distinguish between ‘good’ and ‘bad’ job applicants. These characteristics of labour markets can lead to market distortions that prevent a textbook competitive market outcome.
Recent court developments
Court opinions relating to labour market definition in two ongoing cases support a central role for market definition and illustrate the need to consider both worker characteristics and job attributes.
Deslandes v McDonald’s
In a class action against McDonald’s (Deslandes v McDonald’s), plaintiffs alleged that no-hire agreements in McDonald’s franchise agreements were anticompetitive. The plaintiffs argued that the hiring restraints should be analysed under a per se standard or, alternatively, under a quick-look approach, while the defendants argued that the restraints should be judged under a rule of reason approach. In its opinion denying class certification, the court found that analysis of the no-hire agreement required a rule of reason analysis and, citing Ohio v American Express Co, the court further found that this ‘requires courts to conduct a fact-specific assessment of “market power and market structure”’.
The court found that the ‘relevant market for employees to do the type of work alleged in this case . . . who hold low-skill retail or restaurant jobs . . . is a small, geographic area’. To identify the other jobs, if any, in the relevant market, the court cited evidence of franchisees and other local employers competing for the same workers. In doing so, the court rejected the plaintiffs’ assertion that the relevant market should be the market for McDonald’s restaurant workers, which the court observed would mean limiting the market to a single brand, which ‘defies logic’.
The court ultimately dismissed the remaining claims of the individual plaintiffs, finding that neither plaintiff could plausibly allege that McDonald’s had labour market power in the relevant market. The court again rejected a relevant market limited to McDonald’s restaurants, saying that a relevant market ‘limited to McDonald’s outlets is implausible. They could have sold their labour to other buyers’. The court noted that:
[i]t is undisputed that, within three miles of Deslandes’s home were two McDonald’s restaurants and between 42 and 50 other quick-serve restaurants. Within ten miles of Deslandes’s home were 517 quick-serve restaurants. Accordingly, Deslandes cannot plausibly allege that defendants had market power in the relevant market within which she sold her labour. Within ten miles of Turner’s home were 253 quick-serve restaurants. Accordingly, Turner cannot plausibly allege that defendants had market power in the relevant market in which Turner sold her labour. Without market power, defendants could not suppress plaintiffs’ wages; another buyer would step in to pay plaintiffs more.
Jien v Purdue Farms
In Jien v Purdue Farms, the plaintiffs alleged a conspiracy by poultry processors and several data consulting companies to fix and depress compensation and a conspiracy to unlawfully exchange compensation data. The court determined that the information exchange claim should be evaluated under the rule of reason and that that required an analysis of the relevant market to evaluate anticompetitive effects.
The Jien plaintiffs argued for a relevant market definition limited to jobs in the poultry industry, pointing to worker characteristics such as low educational attainment, limited language skills and industry-specific skills, which they argued limit their ability to transition to other jobs. In contrast, the defendants argued that the relevant market should include other jobs located in the same geography that hire low-skilled workers. In its opinion on motions to dismiss, the court found that the relevant market definition was sufficient for this early stage of the litigation even though some of the allegations ‘will require significant fleshing out to survive challenge at later stages of this litigation’.
Notably, the court found:
[t]he poultry processing labour market is plausible, because Plaintiffs have sufficiently pled facts suggesting the Defendant Processors’ employees are reasonably limited to poultry jobs. Specifically, Plaintiffs claim that at least some of [the] poultry workers have developed industry-specific skills . . . Plaintiffs also claim that limited education and language skills make it difficult for poultry processing workers to transition to other jobs.
That is, the court accepted the argument at the early stage of the litigation that poultry workers might be unable to substitute to other, local, low-skilled jobs because other employers did not want to hire these workers. In Jien, the court highlighted the lack of formal education and language skills. It will be interesting to see whether this conclusion is sustained if or when the case is fully litigated.
Given the increased scrutiny of labour markets by antitrust authorities, we expect more cases in which defining the relevant market for labour is a key point of contention. In labour markets, workers often have limited information on available choices and tend to switch jobs infrequently. Moreover, job changes require agreement between employers and employees, which means that employer decisions are also important for labour market definition. Therefore, defining relevant labour markets can present challenges that are not typically found in defining relevant output markets.
Recent cases have involved markets for lower-skilled workers, which can include more substitutes than the corresponding output markets. However, Jien highlights that courts will judge these points on a case- and fact-specific basis.
 Department of Justice (DOJ) Antitrust Division and Federal Trade Commission, ‘Antitrust Guidance for Human Resource Professionals’, 2016, https://www.justice.gov/atr/file/903511/download; Executive Order No. 14036, Promoting Competition in the American Economy, 86 Fed. Reg. 36987 (9 July 2021), https://www.federalregister.gov/documents/2021/07/14/2021-15069/promoting-competition-in-the-american-economy.
 In 2010, the DOJ brought a case against Adobe, Apple, Intel and Pixar, which had allegedly agreed not to cold-call each other’s employees (U.S. v Adobe Systems et al. (1:10-cv-01629), https://www.justice.gov/atr/case/us-v-adobe-systems-inc-et-al), as well as a separate action against Lucasfilm for its no-poach agreement with Pixar (U.S. v Lucasfilm Ltd. (1:10-cv-02220-RBW ), https://www.justice.gov/atr/case/us-v-lucasfilm-ltd). The DOJ brought another no-poach case against eBay for its agreements with Intuit in 2012 (U.S. v eBay Inc. (12-CV-05869-EJD-PSG), https://www.justice.gov/atr/case/us-v-ebay-inc).
 ‘Generally, a relevant market is one which includes “products or services that are reasonably interchangeable with, as well as identical to, defendant’s product” affected by the rule or regulation being challenged. . . . A similar rule applies to supplier markets such as the employment market in the instant case: the relevant market is one where employment positions are reasonably interchangeable with those offered by defendant.’ NHL Players Ass’n v Plymouth Whalers Hockey Club, 419 F.3d 462, 471–72 (6th Cir. 2005).
 The Department of Treasury refers to monopsony or market power as ‘labour market power’, which it defines as ‘a firm’s power to reduce the compensation it pays to its workers, paying less than an equivalent job would, in a hypothetical perfectly competitive market’. US Department of the Treasury, ‘The State of Labor Market Competition’, 7 March 2022, p. 3. George J Borjas defines a monopsony as a firm facing an upward-sloping labour supply curve and notes that ‘individual firms might have some degree of monopsony power even when there are many firms in the labor market competing for the same type of labor’. George J Borjas, Labor Economics, sixth edition (McGraw Hill, 2013), pp. 188, 192.
 C Scott Hemphill and Nancy L Rose, ‘Mergers that Harm Sellers’, The Yale Law Journal, 2018, 127, 2078–2109; US Department of the Treasury (footnote 4), p. 4.
 See, eg, Counsel of Economic Advisors, ‘Labor Market Monopsony: Trends, Consequences, and Policy Responses’, October 2016, p. 2; US Department of the Treasury (footnote 4), p. 4; George J Borjas (footnote 4), pp. 190–191.
 See US Department of the Treasury (footnote 4), p. 3.
 See Suresh Naidu and Eric A Posner, ‘Labor Monopsony and the Limits of the Law’, Journal of Human Resources, vol. 57, supp. 2022, p. S289; Alan Manning, ‘Imperfect Competition in the Labor Market’, in Orley C Ashenfelter and David Card (eds), Handbook of Labor Economics, vol. 4 (Elsevier, 2011), pp. 973–1041. The Treasury Department notes that ‘[t]here is also increasing recognition that market power may be inherent in the firm–worker relationship’, US Department of the Treasury (footnote 4), p. i. Manning notes that ‘many aspects of labor markets are best analyzed from the perspective that there is some degree of imperfect competition. At its most general, “imperfect competition” should be taken to mean that the employer or worker or both get some rents from an existing employment relationship. If an employer gets rents, then this means that the employer will be worse off if a worker leaves i.e. the marginal product is above the wage and worker replacement is costly. If a worker gets rents then this means that the loss of the current job makes the worker worse off – an identical job cannot be found at zero cost. If labor markets are perfectly competitive then an employer can find any number of equally productive workers at the prevailing market wage so that a worker who left could be costlessly replaced by an identical worker paid the same wage. And a worker who lost their job could immediately find another identical employer paying the same wage so would not suffer losses.’ Alan Manning, op. cit, p. 974. Recruitment costs to employers include the costs of selecting and training new workers, as well as uncertainty about the quality of prospective employees. Switching costs to workers may arise from idiosyncratic differences across jobs, such as personalities of supervisors, working hours, relationships with co-workers and so forth. id., pp. 976–978.
 For example, the DOJ argued that many workers in poultry processing plants were unable to substitute to other, low-skilled jobs because the ‘workers face constraints in finding employment that greatly restrict their job options . . . [and] workers with criminal records, probation status, or lack of high school or college education are often able to work at poultry processing plants even when other jobs are not available to them’. United States of America v Cargill et al., United States District Court for the District of Maryland, Complaint, 25 July 2022, ¶ 61.
 As noted by George J Borjas, ‘upward-sloping supply curves for particular firms may arise even when there are many firms competing for the same workers. In short, many firms in competitive markets could have some degree of monopsony power’ (footnote 4), p. 193. See also Naidu and Posner (footnote 8), pp. S284–85; US Department of Treasury (footnote 4), p. 3.
 James B Rebitzer and Lowell J Taylor, ‘Extrinsic Rewards and Intrinsic Motives: Standard and Behavioral Approaches to Agency and Labor Markets’, in Orley C Ashenfelter and David Card (eds), Handbook of Labor Economics, vol. 4 (Elsevier, 2011), pp. 715–720.
 DOJ Antitrust Division and Federal Trade Commission (footnote 1), p. 2.
 In Deslandes v McDonalds, the court cited case documents that included requests to increase wages citing wages at specific local employers. Memorandum Opinion and Order, Leinani Deslandes, Plaintiff, v McDonald’s USA, LLC, McDonald’s Corporation, and Does 1 through 10, Defendants, No. 17-C-4857, United States District Court for the Northern District of Illinois, Eastern Division, 28 July 2021 (Deslandes Memorandum (2021)), p. 21.
 See US Department of the Treasury (footnote 4), p. 3, which notes, ‘labor markets are very different than some product markets, like commodity markets, where the product is relatively homogenous, and the buyers are usually indifferent to who is selling and vice-versa. In the labor market, both buyers (firms) and sellers (workers) take great interest in their counterpart’s characteristics.’
 DOJ, ‘Transcript of Proceedings at the Public Workshop on Competition in Labor Markets’, 23 September 2019, p. 19. See also David Arnold (footnote 14), p. 8: ‘However, some industries and occupations are very specific and there is considerable mobility across both industries and occupations in the data . . . [Using Current Population Survey data,] conditional on switching jobs, the probability the job transition is within the same 4-digit industry is 36.6 percent. . . . conditional on switching jobs, the probability the job transition is within the same 3-digit occupation is about 34.9 percent, slightly lower than the probability of a within-industry transition. Therefore, regardless of whether industry or occupations are used to construct labor markets, it is important to consider the possibility of significant transitions outside of the proposed labor-market definition.’ The set of jobs that are reasonably substitutable can differ across geographic labour markets. For example, in Deslandes, the court noted that the McDonald’s franchisees said that ‘[t]he main competitors for labor in the Jacksonville market are the other quick service restaurants in the vicinity of our restaurants. Comparatively, the main competitors for labor in the Orlando market are the other quick service restaurants, as well as theme parks, Wal-Mart, Sam’s Club, and Costco.’ Memorandum and Order, Deslandes Memorandum (2021), p. 22.
 See David Arnold (footnote 14), pp. 8, 66.
 The plaintiffs alleged that the no-hire provision in McDonald’s franchise contracts ‘violated the Sherman Antitrust Act and suppressed . . . wages’. The no-hire agreement prohibited hiring current McDonald’s employees or those that had been employed within the previous six months (specifically, the franchise agreement included the following: ‘Franchisee shall not employ or seek to employ any person who is at the time employed by McDonald’s, any of its subsidiaries, or by any person who is at the time operating a McDonald’s restaurant or otherwise induce, directly or indirectly, such person to leave such employment. This paragraph 14 shall not be violated if the person has left the employ of any of the foregoing parties for a period in excess of six months.’ Deslandes Memorandum (2021), p. 2. No-hire agreements, also called no-poach agreements, refer to circumstances when ‘two or more employers agree not to solicit or hire each other’s current or former employees’ (US Department of the Treasury (footnote 4), p. 14). Sometimes these agreements are also referred to as non-solicitation agreements (see, for example, Gibson Dunn, ‘Employers Beware: Aggressive and Expansive Labor-Focused Antitrust Enforcement Will Remain the New Normal’, 18 April 2022, footnote 2, https://www.gibsondunn.com/employers-beware-aggressive-and-expansive-labor-focused-antitrust-enforcement-will-remain-the-new-normal/. Alternatively, non-solicitation agreements can refer to agreements where an employee agrees not to solicit a company’s clients or customers after leaving the company (see US Department of Treasury (footnote 4), p. 14).
 Memorandum Opinion and Order, Deslandes, Plaintiff, v McDonald’s USA, LLC, McDonald’s Corporation, and Does 1 through 10, Defendants, No. 17 C 4857, United States District Court for the Northern District of Illinois, Eastern Division, 25 June 2018, p. 10.
 Ohio v American Express Co., Supreme Court of the United States, 138 S. Ct. 2274, 25 June 2018.
 Deslandes Memorandum (2021), pp. 16–17.
 id., pp. 21, 23.
 id., pp. 21–22. Further, the court noted that whether or not McDonald’s had market power would require analysis of each relevant market, saying: ‘It will undoubtedly be true that in some relevant markets, McDonald’s restaurants will have so many competitors for labour that the restraint will have no anticompetitive effect. In other markets, . . . [McDonald’s restaurants] may have so little outside competition for employees that the restraint will impact the market.’ id., p. 24.
 id., p. 20.
 Memorandum Opinion and Order, Deslandes, Plaintiff, v McDonald’s USA, LLC, McDonald’s Corporation, and Does 1 through 10, Defendants, No. 17 C 4857 and Turner, Plaintiff, v McDonald’s USA, LLC, McDonald’s Corporation, No. 19-C-5524 United States District Court for the Northern District of Illinois, Eastern Division, 28 June 2022, p. 11.
 id., p. 12.
 Jien v Perdue Farms, Inc. et al., Civil Case No. 1:19-CV-2521-SAG, United States District Court for the District of Maryland, 16 September 2020, Memorandum Opinion, pp. 1–2.
 id., p. 20 (‘The parties dispute whether an analysis of the relevant product and geographic markets is required to show anticompetitive effects. . . . the full rule of reason analysis, including definition of the relevant market, is required.’).
 id., pp. 23–24.
 id., pp. 24–25.
 id., p. 25.
 ibid. The court further noted, ‘It is undoubtedly true that the Amended Complaint lumps together a broad range of poultry processing jobs at different pay levels, requiring different skills, and with different job descriptions, and then offers only a handful of details regarding how this diverse array of jobs can be treated together as one single product market separate from other facially similar industries. Nevertheless, only a bare minimum level of specificity is required at this early stage of the proceedings.’