The paper – which can be found here – looks at mergers that facilitate anticompetitive wage and salary suppression from an antitrust perspective. It also looks at other potentially anticompetitive practice in labour markets, so the paper’s title is misleading.
The paper’s fundamental argument is that that antitrust law is under-enforced as regards mergers affecting employment markets, and that this is important for several reasons. First, the share of the gross domestic product (GDP) going to labour has been declining at an alarming rate, and this seems to be correlated with an increase in market concentration. Second, US antitrust law does not condemn unilateral price setting by dominant firms – including the setting of wages. A second best solution to the problem of suppressed wages can therefore be found in merger law, which can interdict wage-suppressing mergers before they occur. Third, antitrust law is properly directed at all output reducing practices, and there is certainly no principled reason for excluding anticompetitive effects in labour markets from antitrust’s scope.
The paper begins by setting the scene – one where no court has ever condemned a merger because of its anticompetitive effects in labour markets. However, some mergers may be unlawful because they injure competition in the labour market by enabling the post-merger firm to suppress salaries to infra-competitive levels. Mergers affecting labour markets require some rethinking of merger policy, but do not challenge this policy’s fundamentals. In a perfectly competitive labour market, each worker would receive the marginal value of her labour – which would reflect her productivity. However, a labour monopsony depresses employment below the level that would obtain in a perfectly competitive market. The employer with market power in the market where it purchases labour will suppress its purchases, driving salaries down. This will lead to higher profits because the employer can pay workers less than their productivity, capturing the surplus for itself. Such an employer will also employ fewer workers if it has market power, since it will be offering lower salaries to all similarly placed jobs. As a result, monopsony leads to both lower employment and lower wages compared to a perfectly competitive labour market. Ceteris paribus, lower employment also entails lower production on the output (product) side.
A subsequent section presents the economic theory and evidence for monopsony in the US labour market. The key message from economic theory is that, as one moves away from the competitive equilibrium towards a situation of monopsony in labour market, wages and production both tend to decrease. Empirical labour economics has studied the topic of monopsony for some time: the key findings from this literature are that monopsony power exists, and that workers are paid below their marginal productivity.
The paper then turns towards assessing those situations that may trigger antitrust enforcement in labour markets.
- One example is anti-poaching agreements – i.e. agreements that occur when employers agree with each other not to hire one another’s workers. Without collusion, firms would normally continue to hire as long as the next employee produced more revenue than the employee would cost the employer. A no-poach agreement enables different firms to act together in the same way as a single firm monopsonist. This, in turn, allows the firms to maximise their own profits by restricting output and paying less than the marginal contribution of each employee – or, in any event, to pay salaries below those that would prevail absent the agreement.
- Another example is non-competition agreements. The difference between a non-competition agreement and a no-poaching agreement is that the former is purely vertical: it refers to agreements between a single employer and its employees. Non-competition agreements are usually justified by concerns with various forms of free riding, such as costly on-the-job training, knowhow or knowledge of trade secrets. As a result, the law of employee non-competition agreements has frequently been perceived as a type of quasi-IP right protection for companies. However, one noticeable and disturbing trend in the market is toward the increased use of non-competition agreements as regards lower-wage and less well trained employees, for whom the quasi-IP rationale is less tenable. Furthermore, employee non-competition agreements may have horizontal effects, particularly if multiple employers in a labour market use them. This may require these agreements to be subject to a horizontal effects’ analysis, either under the rule of reason or for merger control purposes.
For the purpose of merger control in labour markets, the most important question is whether the merger is likely to increase the extent of labour monopsony, thus reducing wages and output. Ceteris paribus, as labour market concentration levels rise, predicted wages decline. Consequently, competition agencies will be justified in making a prima facie case against a merger based on existing concentration levels in the affected labour market, as well as on the extent to which the merger increases labour market concentration. If a prima facie case that a merger is likely to result in an anticompetitive effect is made out, the merging parties may still prove efficiencies to the effect that the merger is procompetitive. To have a chance of succeeding, an efficiency case for a merger affecting a labour market must show that post-merger reorganization will decrease the need for workers but will not lower total production. Both of these requirements are essential. A merger that decreases the need for workers may represent nothing more than an exercise in monopsony power; in that case, the merger will also, ceteris paribus, reduce production. By contrast, a merger that eliminates duplication can reduce the need for workers, but production will not go down. Instead, it should go up to the extent that the post-merger firm has lower costs.
The authors also suggest a market definition for labour markets which was already discussed in the paper reviewed in the post above. This market – comprising a commuting zone (geographic market) by 6-digit Standard Occupational Qualification (SOC) (product market) by quarter (time market) – will be hard to apply outside the USA and need not detain us here, but may be relevant for those with an interest in the topic. The most interesting part of the authors’ analysis is how it tries to adapt the SNIIP test to labour markets by developing a ‘small and significant but non-transitory reduction in wages’ test.
To conclude, the authors argue that the consumer welfare standard and traditional antitrust tools can be applied to labour markets. Monopsonies can sow some confusion, however, because they lead to lower, instead of higher prices. As such, it is important to explain how monopsonies can lead to consumer harm.
In some cases, an exercise of monopsony power in the labour market will evidently harm consumers in the product market. This will typically occur when the post-merger firm has market power on both sides of the market. In that case, exercising market power on the labour side will entail the purchase of less labour. Ceteris paribus, less labour will lead to less output on the product side. If the firm has power on the product side of the market, the result will be higher product prices, and evident consumer harm
But harm to consumer welfare occurs just as certainly if the firm exercises monopsony power in the labour market but does not have market power as regards the product market. In that case, the suppressed purchasing of labour will result in lower product output but not in higher prices, given that the product market is competitive. While the post-merger firm’s lost output will be replaced by the production of competitors, employees in the relevant labour market will suffer antitrust injury from the employer’s suppressed output.
Comment: This is a fairly comprehensive overview of how antitrust can intervene in labour markets. It is not as comprehensive as the paper reviewed in the post below, though.