In recent years there has been an upsurge of interest in competition policy applied to labor markets. But the development of a coherent competition policy for labor markets requires us to confront an old debate: What is the economic foundation of competition law? We argue that both the total welfare and consumer welfare standards are flawed. An often-proposed alternative (the “protection of the competitive process standard”) lacks a foundation in economic welfare principles and therefore provides no guidance in resolving trade-offs. Fortunately, there is an alternative. We suggest that a newly emerging approach, known as the transactions cost approach to competition law, offers promise as a credible alternative foundation for competition policy in labor markets and other antitrust markets more generally.
Darryl Biggar, Allan, Alberto Heimler1
I. INTRODUCTION
In recent years there has been a substantial increase in interest in the application of competition law to labor markets. We believe this is a positive development. Labor markets have, for too long, been seen as off-limits to competition enforcers. As a first step in the reinvigoration of antitrust enforcement in the context of labor markets, several recent papers have sought to articulate a clear competition policy for labor markets.2 In developing a competition policy for labor markets, antitrust economists are forced to confront, once again, the question of the economic goals of competition policy. The application of competition law and policy to the new domain of labor markets forces us to look again at old debates.
Until recently, the conventional wisdom among most competition scholars was that competition law and policy seeks to promote some form of consumer welfare. But in recent years this approach has come under sustained attack. This loss of confidence has arisen, in part, from concerns that an exclusive focus on the downstream or output side of the market (as the consumer welfare standard seems to require) has the potential to overlook anti-competitive harm in upstream or input markets – precisely the domain in which labor markets fall. It doesn’t seem right to begin our analysis of competition policy in labor markets by applying an economic standard which essentially neglects the economic consequences of anti-competitive actions in input markets upstream.
As economists, it seems more natural to us to apply the concept of total welfare or total surplus – the same concept that is applied throughout the economic analysis of public policy. This approach identifies the key economic harm from the exercise of market power as the loss in welfare known as deadweight loss. Over the years many highly regarded economists have argued that competition decision-makers should seek to apply a concept of total economic welfare in all areas of competition policy analysis.3 All of the recent academic papers exploring the application of competition policy to labor markets have implicitly or explicitly adopted a “total welfare” approach.4
However, there is a fundamental problem: It is universally recognized amongst competition scholars that competition authorities do not behave as though the promotion of the traditional concept of total economic welfare is their primary concern.5 If competition agencies do not pursue total economic welfare in their other enforcement activities, why should we expect them to do so in the context of labor markets?
This article suggests that there is another way. For the reasons set out below, we believe that a newly emerging approach – known as the transactions cost approach to competition law – offers promise as a credible economic foundation for competition law in general, and to competition law and policy in labor markets, in particular.
This paper does not seek a full elaboration of competition policy for labor markets based on the transaction cost approach. Such an elaboration is beyond the scope of this short paper. Instead, we merely seek to turn a spotlight on the economic foundation for competition policy, particularly in the context of labor markets.
II. AN ECONOMIC FOUNDATION FOR COMPETITION IN LABOR MARKETS
What, exactly, is the economic foundation for competition policy? This question becomes particularly important in the context of competition policy in labor markets.
A. Consumer Welfare and Input Markets
According to the conventional consumer welfare standard, in assessing whether or not a particular action should be deemed to be anti-competitive (i.e. in breach of the competition law), attention should be paid to the impact on downstream consumers. If consumers are not left worse off, the practice is deemed to be benign, if not actually efficient or pro-competitive. Heyer’s so-called First Theorem of Antitrust can be summarized as follows: “If consumers are hurt by the merger, it is anti-competitive and should be blocked. If consumers benefit, the merger is pro-competitive and should be allowed.”6 According to Herbert Hovenkamp (2019):
Antitrust law in many jurisdictions defines its consumer welfare goal in terms of low consumer prices. For example, mergers are challenged when they threaten to cause a price increase from reduced competition in the post-merger market. While the consumer welfare goal is under attack in some circles, it remains the most widely expressed goal of antitrust policy in the United States.7
Despite its popularity, the consumer welfare standard has several, fundamental flaws. First, and perhaps foremost, the consumer welfare standard, with its exclusive focus on the buyer side of the market, seems particularly ill-suited to the analysis of competition issues in input (or supplier) markets. On a strict consumer welfare standard, market power exercised by dominant buyers in labor markets (or other input markets) is at worst benign, and is possibly beneficial, for downstream consumers. This is a problem for scholars who seek to develop competition principles in labor markets.8
Indeed, as Naidu, Posner and Weyl (2018) point out, the exercise of market power by a dominant buyer in labor markets might be considered pro-competitive: “wage suppression even as it hurts workers, at least benefits consumers.”9 Hovenkamp (2019) observes that anti-competitive restraints in labor markets might result in lower prices to consumers, which should be promoted under the consumer welfare standard:
Focusing entirely on [output] price makes it awkward to work the supply side of markets into debates about consumer welfare. Labor markets are a notable example. Labor appears in the market as suppliers, not as purchasers. While consumers-as-consumers benefit from lower prices, combatting restraints in labor markets generally focuses on wage suppression. That is, today the principal problem of competition policy in labor markets is wages that are too low, not those that are too high.
In seeking to develop a sound competition policy for the labor market, the consumer welfare standard is of little use. We cannot develop sound competition principles for labor markets by starting with an economic approach which asserts that the exercise of market power by a dominant firm in labor markets is benign or even pro-competitive.10
In any case, competition enforcers do not, in practice, behave as though “consumer welfare” is their primary concern. Many antitrust scholars have pointed out that a literal application of the consumer welfare standard would exempt anti-competitive practices on the buyer side of the market, such as merger to monopsony, as long as those practices did not result in harm to consumers.11 This is inconsistent with the way that most antitrust laws are drafted and in the way the majority of competition enforcers around the world behave. In practice, competition law “is applicable to monopsony power even where there is no harm to consumers downstream, under analogous legal standards as those of product markets.”12
B. Total Welfare as the Foundation of Competition Policy
Is there an alternative to the consumer welfare standard? As economists, the most natural economic foundation for competition policy is the same welfarist foundation that is used throughout all of public policy analysis – the concept of total welfare. Many highly respected economists have argued that, when it comes to competition policy, governments and courts should seek to maximize a concept of total economic welfare, usually reflected in the sum of producers’ surplus and consumers’ surplus.
Hovenkamp, for example, argues that competition policy should focus exclusively on output, not price:
We would do better … to define [the goal of competition law] in terms of output rather than price. Competition policy should strive to facilitate the highest output in any market that is consistent with sustainable competition. … When ‘consumer welfare’ is defined in terms of output it becomes much easier to articulate a defensible competition policy that does everything that antitrust can properly do to ensure a healthy economy, reflecting both the buy and sell sides of the market.13
Hovenkamp’s focus on output, rather than price, is a direct consequence of the total welfare standard. Hovenkamp is correct that, if competition authorities are serious about pursuing a total welfare standard, they should focus exclusively on sales volumes and not on price.
Unfortunately, it is universally agreed amongst competition economists that competition enforcers do not pursue the textbook concept of total economic welfare. For example, Heyer concludes:
Neither the US competition authorities, nor competition authorities in most other economies, appear willing to adopt explicitly and unambiguously a total welfare standard for merger enforcement.14
Farrell & Katz (2006) go further to argue that competition authorities routinely ignore the total welfare implications of the economic models they use. They conclude: “Evidently, either we don’t trust those models, or we don’t believe in a purely welfarist total surplus standard.”15
In addition, the competition laws themselves, especially those that follow the EU model, include examples of possible violations which are difficult to justify if promoting conventional total welfare is their primary concern.
Nonetheless, the total welfare standard has been adopted in the recent papers arguing for stricter competition law enforcement in labor markets. For example, Naidu, Posner and Weyl (2018):
monopsony power … 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 non-employment as workers find prevailing wages unacceptable and exit the labor force or refuse to take available jobs.
Despite its adoption by Hovenkamp, Naidu et al., and others, is the total welfare standard a sound foundation for competition policy in labor markets? We believe there are reasons for concern.
The first reason for concern is that, due to inelasticity of labor supply and/or the ability of labor hirers to price discriminate, the economic welfare loss from market power in the labor market (as conventionally measured) may simply be too small to justify intervention. For those who believe in the need for re-invigoration of antitrust enforcement in labor markets, this is a serious intellectual hurdle to overcome.
For example, Krueger16 emphasizes that the overall elasticity of labor supply is very low – any given change in wages results in a very small reduction in employment. It follows that the deadweight loss from a reduction in wages is likely to be small.
Even if the elasticity of labor supply is material, the impact of market power on employment levels may be insignificant if the labor-hirer is able to price discriminate. The potential for price discrimination is far greater in labor markets than in most product markets. In most consumer-product markets, prices are public and posted before the seller knows anything about the identity of the buyer. In contrast, wages and salaries are typically highly confidential information. Individual workers may have an idea of the range of salaries for work of a particular category, but workplace taboos make it difficult to know exactly what other workers are paid. This materially increases the scope for the labor-hirer to engage in wage discrimination.
Furthermore, many employers pursue a practice of not revising wages until the employee expresses a willingness to take up an outside offer. At this point the employer (who has a lot of information about the employee, including their willingness and ability to relocate to find alternative employment) may match the outside offer. This practice ensures that each employee is paid just enough to keep them in the present employment – a form of “perfect wage discrimination.” According to Naidu, Posner and Weyl (2018) “the emergence of sophisticated prediction algorithms applied to vast troves of human resource data suggest that the ability to wage discriminate in the future may be expanded.”
If the labor hirer can discriminate effectively, the deadweight loss from the exercise of market power in the labor market is likely to be small. The total welfare foundation, with its primary focus on deadweight loss, provides, at best, a mediocre grounds for intervention in labor markets.
C. The Protection of the Competitive Process
Are there alternatives to the total welfare or consumer welfare standards? In recent years, there has been increasing support for the proposition that the primary purpose of competition law is the protection of the process of competition (see, for example, Khan17, and Wu18). This approach has a long historical tradition and has recently received surprising support from such distinguished authorities as Carl Shapiro19 and Renata Hesse20.
We are sympathetic to, and respect, those who advocate for the “protection of the competitive process” standard. We acknowledge that much of the day-to-day work of competition law enforcers can loosely be described as “protecting the competitive process.”
But, the protection-of-the-competitive-process standard lacks a foundation in conventional economic welfare concepts. As a consequence, this standard cannot make clear predictions in circumstances where competing economic objectives are placed in conflict. Intervening in a market will often involve trade-offs between the interests of different market participants, or between conflicting objectives, such as tradeoffs between innovation and productive efficiency, or between incentives for upstream innovation and downstream investment. The competitive process standard is silent on such trade-offs. As FTC Commissioner Caroline Wilson has pointed out, this leads to the risk of unpredictability and inconsistency.21
III. THE TRANSACTION COST APPROACH TO COMPETITION LAW
An alternative economic foundation for competition law has recently been proposed.22 This approach, known as the transaction cost approach to competition law, argues that competition law exists to protect, and thereby promote, sunk, relationship-specific investments in the economy.
This approach starts from the observation that nearly all on-going economic relationships will require some form of sunk investment by one or both of the transacting parties. Such sunk investments are well known in labor markets. Most workers spend years developing general human capital (e.g. a general education) together with a wide range of special skills (e.g. as a chef, or carpenter). These investments are made in the expectation that they will raise the wage of the labor-supplier in the long-run.
But the sunk investment of labor suppliers is not limited to human capital. In most cases the supply of labor requires the physical presence of the labor supplier. As a result, most labor markets are strictly local (within commuting distance). Most workers will therefore make a significant sunk investment in a location. For example, a worker may choose to relocate her family to find a job in a small town. The worker and her family take some time to get established in the community – to find a job for her spouse, to establish relationships with neighbors, to develop local service relationships (e.g. a local doctor or dentist), to get the kids set up in schools and so on. The time and cost required to establish these arrangements is a form of sunk investment in reliance on the expectation of employment in the town.
Like all sunk investments, these sunk investments face the threat of hold-up – the threat that, once sunk, the other party to the transaction will seek to change the terms and conditions of trade.23 In the context of labor markets, the threat of hold-up is the threat that, once the worker has become established in a job the labor-hirer will cut the wages (or, equivalently, refuse a promised increase in wages).
Fortunately, competitive markets offer some protection against the threat of hold-up. In a competitive market, the investment by one side of the market is not specific to one particular trading partner. In the event of an attempt at hold-up, the party who has made the sunk investment can switch to another trading partner. If there is more than one suitable potential employer within commuting distance, a worker faces at least some protection against hold-up: In the event of a threat to cut in wages the worker can, in principle, switch to an alternative employer.
In addition to competitive markets, certain private arrangements can offer some protection against the threat of hold-up. The most common example is a simple contract. A worker who must make a substantial sunk investment in relocating his/her family may seek a long-term employment contract, rather than risk that, after relocating, a promised pay rise will not materialize. Another possible mechanism for protecting against hold-up is vertical integration as in the case of worker-owned co-operatives.
But vertical integration and long-term contracts are not always feasible. Almost by definition, relatively few workers are in a position to purchase the company that employs them. Very few companies are organized as worker co-operatives. It is theoretically possible to envisage a labor contract which specifies the terms and conditions of the labor supply that would apply in the event of a change in the market structure (e.g. a merger, or insolvency) that changed the outside options for the worker. But, in practice, such contracts are infeasible. Furthermore, theory suggests that the life of the contract must be as long as the life of the underlying sunk investment. But some sunk investments may last decades (e.g.
the human capital investment by a doctor in training). Long-term labor supply contracts with a life of decades are almost completely infeasible.
As a consequence, circumstances can arise where in an on-going economic relationship the sunk investment of one side is subject to the threat of hold-up. In the absence of government intervention, this could have a chilling effect on investment. Valuable economic activity would be foregone. For example:
- Workers may be unwilling to move to a “company town” with a single employer, even though their productivity would be higher in that location, due to the threat that, once there, wages would not increase as promised;
- Individuals would be unwilling to invest in new skills or human capital out of fear that there would be no increase in wages to justify the cost of the investment.
Conversely, on the other side of the market, in circumstances where it is the labor suppliers that have market power (either individually or collectively), labor-hirers (firms) may be unwilling to invest in more productive equipment, out of fear that, once the investment is sunk, the workers will hold-out for a bigger share of the firm’s profit.
According to the transaction cost approach to competition law, a transacting party exercises market power when it engages in hold-up; that is, when it seeks to change the terms and conditions of trade. The economic harm from the exercise of market power is the chilling effect this has on incentives for socially-valuable investment.
The transaction cost approach to competition law suggests that competition law steps in to protect, and thereby promote, the sunk investment of transacting parties. Specifically, competition law imposes two types of controls:
(a) In the case of parties in a dominant position (i.e. a firm in a position to engage in hold-up) competition law imposes strict limits on its ability to engage in hold-up – including limits on its ability to price discriminate, to change its prices, or on its ability to withhold services; and
(b) In the case of parties which might collectively be in a dominant position, through merger or collective arrangements, competition law imposes strict limits on their ability to merge or collude to put them in a position to engage in hold-up.
IV. THE BENEFIT OF THE TRANSACTION COST APPROACH TO COMPETITION LAW
In our view, the transactions cost approach to competition law is valuable for two reasons: (a) it is based in conventional economic total welfare or total surplus principles; and (b) it can help understand and explain the behavior of competition decision makers in situations where the total welfare and the consumer welfare standards are lacking.
As just noted, the transactions cost approach is based on total welfare or total surplus principles. The key difference to the way in which the total welfare approach is normally applied is that the transactions cost approach allows for the transacting parties to make sunk investments which shift their supply or demand curves. The resulting change in welfare can easily exceed (by many orders of magnitude) the economic harm due to the deadweight loss. In this sense the transactions cost approach allows for a more dynamic framework compared to the “static” model within which the total welfare concept is usually applied.
In our view, this small change in the economic model leads to predictions which are more consistent with competition laws and competition decisions in practice. For example, under the conventional total welfare standard, we should expect to see competition authorities approve mergers which enhance the ability of parties to price discriminate. In contrast, under the transactions cost approach, price discrimination – by allowing the dominant firm to charge different prices depending on the degree of reliance by its customers on its services – will allow the dominant firm to surgically expropriate some of the value of the sunk investment, reducing the incentive to invest. Rather than actively encouraging price discrimination, the transactions cost approach to competition policy suggests that we should expect to see competition authorities taking a dim view of price discrimination, consistent with what we observe in practice.
Similarly, we have seen that, under the conventional total welfare standard, competition authorities should turn a blind eye to mergers where demand is inelastic (downstream) or supply is inelastic (upstream). The transactions cost approach to competition law suggests that it is precisely when demand or supply is inelastic that the sunk investments of trading partners are most exposed to the risk of being expropriated.
In the same way, it is common for competition authorities to take action to control the behavior of firms in a dominant position – whether in a dominant position in an output market, or in an input market (such as a labor market). Under the conventional total welfare approach, the justification for such intervention is limited to situations in which rivals in an upstream or downstream market are excluded from the market.24 In contrast, under the transactions cost approach, actions by a dominant firm to expropriate the sunk investment of trading partners should be prohibited (whether market participants are excluded or not). This might include engaging in price discrimination, or reducing the prices to longstanding input suppliers, such as labor-suppliers.
We can also contrast the transactions cost approach with the consumer welfare standard. As noted earlier, a strict application of the consumer welfare standard might require a competition authority to turn a blind eye to anticompetitive behavior upstream (such as in the labor market) particularly when such behavior might allow the firm to lower prices to consumers downstream. Under the transactions cost approach however, as long as the upstream suppliers have made material sunk investments, the transactions cost approach suggests that competition authorities should prevent firms from taking advantage of their position to expropriate the value of investments upstream. This might include, for example, both the case of a dominant employer lowering wages to captive employees, or a dominant on-line platform taking advantage of its position to exploit the sunk investment of its suppliers.
Finally, we consider that the transactions cost approach offers benefits over the “protection of the competitive process” standard. For example, there could in principle arise trade-offs between the need to protect sunk investment by trading partners, or sunk investment by the dominant firm itself. Such trade-offs have no resolution under a “protection of the competitive process” standard. The transactions cost approach allows us to continue to view competition not as an end in itself, but as a means to an end – an end which we can continue to assess through a conventional economic welfare lens.
In summary, we suggest that, alongside the range of other labor market regulations, competition law has a role to play in protecting workers from the threat of hold-up by labor hirers. A full articulation of the implications of the transactions cost approach to labor market competition policy is beyond the scope of this paper. We consider that this approach offers credible foundation for competition policy in the context of labor markets. More generally, in our view, the transactions cost approach to competition law offers promise as a credible economic foundation in many other antitrust markets in the future.
1Respectively, Special Economic Advisor, Australian Competition and Consumer Commission (“ACCC”), Professorial Fellow, Melbourne Law School, and former chairman of the ACCC; Professor of Economics, Italian National School of Government, (Scuola Nazionale dell’Amministrazione), Rome, Italy, and Chairman, Working Party 2 of the OECD Competition Committee. The views expressed are the authors’ own.
2 See, e.g. Naidu, Suresh, Eric A. Posner, E. Glen Weyl, (2018), “Antitrust Remedies for Labor Market Power,” 132 Harvard Law Review 536; Azar, José, Ioana Marinescu & Marshall Steinbaum, (2019), “Antitrust and Labor Market Power,” Research Brief, Economics for Inclusive Prosperity, May 2019.
3 See, e.g. Farrell, Joseph & Michael Katz, (2006), “The Economics of Welfare Standards in Antitrust,” 2 Competition Policy International 3, 6-7; Carlton, Dennis W., (2007), “Does Antitrust Need to be Modernized?,” 21 Journal of Economic Perspectives 155-176.
4 See Naidu, Posner & Weyl (2018) supra note 2; Hovenkamp, Herbert, (2019), “Competition Policy for Labour Markets – Note by Herbert Hovenkamp,” OECD Roundtable on Competition Issues in Labour Markets, June 5, 2019.
5 Heyer, Ken, (2012), “Welfare Standards and Merger Analysis: Why Not the Best?,” (2012) 8 Competition Policy International 146-172; Farrell & Katz (2006) supra note 3.
6 Heyer (2012) supra note 5.
7 Hovenkamp (2019) supra note 4.
8 This is a key part of the critique of Hipster Antitrust. See Khan, Lina, (2017), “Amazon’s Antitrust Paradox,” 126 Yale Law Journal 710-805.
9 See also Azar, Marinescu & Steinbaum (2019) supra note 2.
10 In addition, the consumer welfare standard is inconsistent with standard economic welfare principles which take into account the welfare of both sides of a market (producers and consumers).
11 Carlton (2007) supra note 3, page 158: “if only consumers matter, then a buying cartel should be perfectly legal and indeed should be encouraged.”; Kaplow Louis & Carl Shapiro, (2007), “Antitrust,” in A. Mitchell Polinsky & Steven Shavell, Handbook of Law and Economics (2007), page 1168: “If only consumer welfare mattered, increases in buyer power through horizontal mergers and otherwise might be praised, not condemned.”
12 OECD (2019), “Competition Concerns in Labour Markets – Background Note,” DAF/COMP(2019)2, 5 June 2019, page 34. Hovenkamp (2019) supra note 4, page 6 emphasises that the merger provisions of the Clayton Act “apply to any merger whose effects may be substantially to lessen competition or create a monopoly in any line of commerce, not distinguishing buyer from seller effects.” 13 Hovenkamp (2019) supra note 4.
13 Hovenkamp (2019) supra note 4.
14 Heyer (2012) supra note 5, page 154. See also Lande, Robert H., (1989), “Chicago’s False Foundation: Wealth Transfers (Not Just Efficiency) Should Guide Antitrust,” 58 Antitrust Law Journal 631 (1989).
15 Naidu, Posner & Weyl (2018) supra note 2 seem to want to have a bet each way, although their paper is explicitly based in the conventional economic notion of total economic welfare, they abandon this approach at a key point, arguing that a merger which lowers wages to workers should be prevented, regardless of the impact on total economic welfare.
16 Krueger, Alan B., (2018), “Reflections on Dwindling Worker Bargaining Power and Monetary Policy,” Luncheon Address at the Jackson Hole Economic Symposium, August 24, 2018.
17 Supra note 8.
18 Wu, Tim, (2018), “The ‘Protection of the Competitive Process’ Standard,” Federal Trade Commission Hearings, November 1, 2018.
19 Shapiro, Carl, (2017), “The Consumer Welfare Standard in Antitrust: Outdated, or a Harbor in a Sea of Doubt?,” statement before the Senate Judiciary Committee, December 13, 2017.
20 Hesse, Renata, (2016), “And Never the Twain Shall Meet? Connecting Popular and Professional Visions for Antitrust Enforcement,” Speech to 2016 Global Antitrust Enforcement Symposium, Washington DC, September 20, 2016.
21 Wilson, Christine, (2019), “Welfare Standards Underlying Antitrust Enforcement: What You Measure is What You Get,” Address at George Mason Law Review Symposium, February 15, 2019.
22 See Biggar, Darryl & Alberto Heimler, (2019), “Re-Thinking The Foundation of Antitrust Law in the Era of Hipster Antitrust: Is Protecting Sunk Investments a Primary Economic Rationale for Antitrust Law?,” under review at Journal of Antitrust Enforcement.
23 The role of sunk investments in labour markets is highlighted in a number of papers. For example, Wachter, Michael L. & George M. Cohen, (1987), “Law and Economics of Collective Bargaining: An Introduction and Application to the Problems of Subcontracting, Partial Closure and Relocation,” 136 University of Pennsylvania Law Review 1349: “Workers make sunk investments in their jobs by agreeing to long-term implicit contracts that provide for ‘deferred compensation’, that is, below-market wages at early stages of employment and above-market wages at later stages. … [W]orkers invest in job-specific training and deferred compensation represents the quasi-rents earned on this training. … Once workers make a sunk investment in monitoring or job-specific training the firm acquires bargaining power over them because it can strategically force the workers to suffer a sunk cost loss by misrepresenting product market conditions, thereby expropriating the workers’ expected return.”
24 Here we are putting aside the “exploitative” abuse of a dominant position which is present in EU, but not U.S. or Australian competition law. But even in the case of an exploitative abuse the total welfare standard says that the dominant firm should be able to charge what it likes provided it does not reduce total sales volumes.
