Redirecting antitrust enforcement to confront monopsony power would be a substantial departure from the way it has been conducted in recent decades, but just because a policy has been in place for a long time does not mean it is a success, and recent evidence implies a significant policy change is necessary and justified.
The very first sentence of the United States submission to the OECD’s “Global Forum on Competition” in 2015 reads “The U.S. Federal Trade Commission and the Antitrust Division of the Department of Justice do not consider employment or other non-competition factors in their antitrust analysis.”
Employment isn’t a “competition factor?” A growing body of evidence, drawn from both micro and macro sources, implies that the labor market has been slack since the Great Recession thanks to an aggregate demand shortfall. The employment-to-population ratio, labor force participation, job-to-job and geographic mobility, and the job ladder as a whole have all stagnated for almost a decade, and even before 2008 these indicators had barely made up ground since the recession of the early 2000s. Wages have been stagnant, and thus the labor share of national income has been in decline.
Over an even longer period, worker compensation has failed to keep pace with productivity gains per capita. The broadening literature on the rise of earnings inequality between firms, controlling for worker characteristics, implies that workers do not receive sufficient job offers to equalize the earnings they receive across firms. All of these phenomena are explained by rising monopsony power in the labor market.
The recent paper by De Loecker and Eeckhout makes this point explicit: the ability of employers to extract rents in the labor market causes a reduction in the market demand for labor, and this in turn motivates all of the manifestations of a slack labor market just described, as well as the rising markups that are the central finding of that paper. Profits have risen, reflecting rising market power in both labor and product markets—a finding that is consistent with the aggregate analysis done by Barkai.
So if monopsony power constrains employment, why not consider it a factor in antitrust analysis? After all, the premise of economic analysis in antitrust is that market power threatens welfare by restricting output and raising prices. Why doesn’t market power threaten welfare by reducing demand for labor and lowering wages?
Existing antitrust policy treats maximizing consumer welfare as the ultimate end goal of antitrust policy, and that policy aim makes sense in a world of little market power, where profits are low and the economy is assumed to be on its production possibility frontier. In that world, maximizing consumer welfare is a suitable summary statistic for overall wellbeing.
Moreover, since the revolution in antitrust policy associated with Robert Bork put such an emphasis on sustaining an “economies” defense—meaning that potentially efficiency-enhancing aspects of corporate mergers and conduct must be weighed against inefficiencies arising from market power—the potential for monopsony power has been considered a plus. After all, enforcers and courts generally assumed that any gains at the expense of workers would be passed to consumers in the form of lower prices, since product markets would be competitive (and if not, the potential for entry would exert a disciplining influence). In that world, caring about employment, wages, or other labor market outcomes looks like protectionism, impeding the competitive pressure that yields the best outcomes for consumers, and favoring certain privileged “insider” workers at the expense of others.
We know now that we don’t live in that world, and that revelation calls for a wholesale re-think of the proper goals of antitrust policy, very much including whether the sole focus on consumer welfare makes sense when powerful corporations squeeze workers and then pocket the gains for themselves and their shareholders.
What would it actually look like to bring antitrust into the labor market?
As with any enforcement regime, antitrust often starts with the lowest-hanging fruit: out-and-out written evidence of anti-competitive practices, such as the Justice Department’s 2010 lawsuit against Silicon Valley employers for colluding not to hire one another’s programmers. This is partly why the recent increase in the use of non-compete clauses has drawn attention in antitrust circles. As a would-be vertical restraint, non-compete clauses aren’t as easy to target under antitrust as horizontal collusion, but they are there, in writing—prohibitions on competition in the labor market, to the benefit of employers. And they should be banned, or at the very least subjected to a high burden of proof requiring a substantive defense on the part of employers who impose them, plus an affirmative finding that they do not act to reduce wages or restrict job offers.
The same dynamic is at play in prohibitions on poaching in franchising agreements, which Alan Krueger and Orley Ashenfelter recently found to be prevalent in franchising contracts and which, to my knowledge, the federal competition regulators have never touched—even though they do regulate other provisions of those contracts. Franchising networks are a hybrid beast, somewhere between horizontal and vertical, but a blanket prohibition on poaching throughout a franchisor’s network certainly starts to look like a horizontal agreement not to compete.
It’s important to understand, though, that these written restraints of trade are symptoms of the broader decline in worker power, and meaningful antitrust enforcement should go after the causes. Reclassification of workers as independent contractors is a broader concern—not only anti-competitive in itself, but as a means to engage in other coercive conduct and corporate structures. Studies show that reclassifications result in immediate wage reductions and no other changes in terms of employment, suggesting that they amount to employer’s exploiting their wage-setting power by changing the legal structure of their business.
And beyond the act itself, classifying workers as independent contractors allows employers to avoid liability for minimum wage, maximum hours, workplace safety, and a host of other entitlements associated with statutory employment. The idea was that employment inherently signifies control, and with control ought to come responsibility—and by extension, if employers do not bear responsibility, then they should not be able to exercise control. What employers have realized now, as enforcement regimes in both labor and antitrust have weakened, is that they can have the control without the responsibility. For example, contracting terms often prevent workers from simultaneously working for others—an exercise of control if ever there was one, and an anti-competitive vertical restraint in the context of an independent contractor.
Employers can have that control without first establishing themselves as a monopoly—in fact, reclassification is increasingly standard operating procedure in many industries, which means that treating it as a violation of Section 2 of the Sherman Act should not require that outright monopolization must first be shown.
This is the fundamental issue behind the litigation over whether Uber’s drivers ought to be considered employees, and if not, whether the business amounts to a price-fixing conspiracy between the company and hundreds of thousands of independent businesses who drive for it. I’ve written before about the antitrust lawsuit against Uber on these grounds. The case was recently dealt a severe blow in the form of an appellate ruling that upheld the company’s mandatory arbitration clause—meaning that if the lower court decides Uber did not itself void the arbitration clause by hastening the case with a move to summary judgment, then the case will likely be thrown out of court.
That brings us to yet another way in which employers exercise monopsony power: mandatory arbitration and class action waivers for employment claims, about which the Supreme Court is set to hear a case this term. The issue there is that expansive readings of the Federal Arbitration Act have essentially said that individual rights protected by both the constitution and federal statute can be voided by bilateral waivers—as though the parties are equally situated and at liberty to reject such provisions in employment agreements and elsewhere. Of course, the whole point of monopsony is that jobs are scarce, and hence employers have leverage with which to extract concessions, be they out-and-out wage reductions or agreements not to litigate disputes. Thus, another aim of antitrust enforcement in labor markets ought to be bans on litigation waivers between parties that are not similarly situated economically as restraints of trade—and the competition authorities ought to make their views known to the Supreme Court on this issue. After all, private action is a pillar of federal antitrust policy, and so arbitration clauses are not just themselves restraints of trade, but they also inhibit enforcement against other restraints, as the fate of the Uber antitrust case shows.
Finally, there’s the elephant in the room when it comes to antitrust: merger review, the bulk of what the agencies do about enforcing the laws they’re entrusted to carry out. Claims that mergers reduce employment are not entertained as arguments against them—in fact, they are likely to be considered arguments in favor, as they show some motivation for the transaction beyond raising prices for consumers. And yet we know anecdotally that recently-consummated mergers have in fact had adverse employment impacts. A systematic study of the labor market impact of past mergers has yet to be conducted, to my knowledge—such an exercise would be a valuable component of assessing the impact and success of the current competition policy regime, including whether these job losses do actually end up benefiting consumers in the form of lower prices, as merging parties invariably claim, versus their shareholders and executives.
In conclusion, the view that the competition authorities expressed to the OECD in 2015 looks increasingly out of touch with the labor market and the broader macroeconomic conditions that currently exist. It is true that redirecting antitrust enforcement to confront monopsony power would be a substantial departure from the way it has been conducted in recent decades, and as such there are both court decisions and agency policies that go against it. But just because a policy has been in place for a long time does not mean it is a success, and recent evidence implies a significant policy change is necessary and justified—much as an intellectual movement in academia once shifted antitrust policy substantially, it’s time for new evidence to change it once again.