The Limits of Antitrust Enforcement

Just a few years ago, the mainstream economic and legal opinions held that labor markets basically tended to “clear,” and that employers’ market power did not have a significant impact on wages or other terms of employment. Today, however, the issue of economic power, especially in the labor market, has made a startling reversal. A number of economists and legal scholars have recently argued that declining antitrust enforcement has harmed not just consumers, but also workers. Employers’ use of non-compete agreements and compulsory arbitration have also come under public scrutiny: non-competes because they limit workers’ ability to leave their employment, and arbitration because it prevents workers from protesting unfair treatment and enforcing their own legal rights. 

Left and right seem to be converging here. Progressives are concerned that corporate power threatens equality, conservatives are concerned that it threatens individual liberty, and both are concerned that it threatens innovation. A populist critique of corporate power run amok may also be good politics. Political culture in the United States has never abandoned the Jeffersonian ideal of the yeoman farmer or independent artisan, nor has it abandoned its characteristic distrust of major institutions. There is now a Congressional Antitrust Caucus, and numerous foundations are sponsoring research into the causes and consequence of market concentration. This is all part of a renewed and essential focus on structural inequality and generally for the good.

I nevertheless want to sound a note of caution. More aggressive antitrust enforcement probably won’t do all that much to help workers, since the problem of employer power runs much deeper than monopsony, covenants, and other restraints on workers’ mobility. Capacious employer rights are written into the basic structure of our labor and employment laws and corporate laws. Those laws encourage investors to aggregate into corporations, while leaving workers atomized, and therefore largely powerless, unless the state encourages them to aggregate into unions or otherwise to exert countervailing power from below. Still more troubling, the basic logic of antitrust—that combinations in restraint of trade are forbidden—is in serious tension with workers’ organizing. Unions are literally cartels for the sale of labor, and unions’ major legal battle prior to the New Deal was to stop courts’ use of antitrust and related doctrines to thwart their efforts. Contemporary antitrust doctrine still has that effect in many instances. 

This all points to a broader tension on the left: on the one hand, many are intuitively attracted to republican or progressive commitments to widely dispersed political and economic power; on the other hand, practically speaking, that may require strong countervailing institutions such as unions that have their own governance authority. Antitrust is clearly part of the solution here, but fundamental labor law reform is equally essential.

•••

Let’s start with the big picture and then drill down. It seems clear that many economic sectors have become more concentrated over the last few decades, which can lead to various social harms. Market power can enable companies to drive up prices for goods; to generate and sustain supracompetitive profits; to prevent market entry by competitors; and to limit innovation. A powerful and aggressive financial sector will encourage such efforts, as they ensure a steady return on capital. To be clear, many antitrust economists would argue that product markets have not become sufficiently concentrated to trigger antitrust scrutiny. But labor markets appear to be a different story.

The economic landscape in many regions has transformed as hospitals have merged, mega-retailers have driven out small businesses, and national fast food chains have replaced local restaurants or regional chains. Now, as two recent papers have argued, many if not most local labor markets in the United States are highly concentrated. Two other recent papers have argued that labor market concentration—whether local or national—correlates with lower wages. Indeed, some are now arguing that workers rather than consumers are the main victims of market concentration today.

This research may help explain a major puzzle in today’s economy: why wages have failed to keep pace with productivity since about 1980. There are at least two plausible causal stories. First, when there are relatively few employers in a field, they may be able to set wages below workers’ marginal productivity. That can involve active collusion. Franchisors have apparently banned franchisees from hiring each other’s workers, for example, and an unlawful non-raiding agreement among various tech firms thwarted workers’ mobility in Silicon Valley for a number of years. Or a handful of employers’ market dominance could enable them to collude tacitly by monitoring one-another’s wage rates. Or the simple fact that workers don’t have many alternative employers may put them at a bargaining disadvantage. In an unpublished talk, economist Heidi Shierholz of the Economic Policy Institute calls such labor market concentration “literal” monopsony. 

Even without literal monopsony, though, workers may find it hard to move among jobs, generating what Shierholz terms “dynamic monopsony.” For example, workers’ skills may be specific to one employer, or they might have very limited information about other job opportunities. Workers also find it quite difficult to relocate for work. With roots in a community and children in a school, workers can’t move as easily as can diversified shareholders and investors. Many firms also restrict workers’ mobility through covenants not to compete. Such instruments, which prevent workers from moving to a firm in the same industry for some period of time, have proliferated even among fast food workers in recent years. They are quite often unenforceable, especially against less skilled workers, and not enforceable at all in California. But the very threat of enforcement can deter workers from leaving or demanding better treatment.

Scholars on both the left and right have begun to argue that employer power is a major problem, and that part of the solution lies in an antitrust revival. Many left-leaning thinkers in this vein—including legal scholars Sabeel Rahman, Tim Wu, and Lina Khan—seek to recast antitrust law in the Progressive-era vein, as a means not just of protecting consumers, but of limiting concentrated power in our political economy. That tradition is perhaps best exemplified by the work of Supreme Court Justice Louis Brandeis, who cautioned that the enormous trusts of the early twentieth century had become states within the state and famously defended Florida’s right to limit the growth of chain stores in order to protect autonomous local businesses. 

The Progressive vision of antitrust waned over the years, in part due to the decline of radical democratic approaches in the late New Deal period and the concomitant growth of the legalistic regulatory and administrative state that tended less to matters of structural economic power. But the death blow for progressive antitrust came from Robert Bork and various other Chicago-school types in the 1970s and 1980s, who argued that mergers tended to enhance consumer welfare and should be scrutinized less closely and that vertical tie-ups should be subject to much less scrutiny.

Were a consumer welfare standard enforced vigorously, it could police cable companies, pharmacies, airlines, automakers, and other providers of basic consumer goods. But it is not enforced vigorously, nor does it contemplate antitrust harms to workers, to possible market entrants, or to political democracy. As with so many Chicago-driven policies, the result has been not to actually open up markets to competition, but to encourage consolidation of wealth and power in fewer hands.

The goals of more conservative antitrust revivalists are not entirely clear, but they seem consistent with undoing parts of the Bork revolution and reinstituting more of a classic liberal economy. Concentrated corporate power is a labor problem, in this view, if it prevents workers from competing fairly in labor markets. Occupational licensing regimes are a problem, apparently, for similar reasons: they limit workers’ ability to enter new professions or to cross state lines. I am extremely skeptical on this latter point, since those regimes often help ensure health and safety and tend to increase workers’ wages. But conservatives’ focus on licensing helps illuminate their basic view of the problem: that workers should be liberated to compete in labor markets.

•••

While such political convergence is tempting, progressives may want to tread carefully here since there is a deeply-rooted tension between antitrust and worker power. Our antitrust laws were written in the late nineteenth and early twentieth-centuries, after the emergence of national-scale enterprises known at the time as “trusts.” The Sherman Antitrust Act of 1890 accordingly declared illegal “every contract, combination . . . or conspiracy, in restraint of trade or commerce.”

The “labor problem” was also an acute political and legal issue at the time—in fact, these were two sides of the same coin. Modern manufacturing corporations had often grown by absorbing smaller firms and then reorganizing production in ways that stripped away workers’ main source of power: their skills. The factory system drove down wages, subjected workers to intense pressure to perform quickly, and led to a massive wave of debilitating industrial accidents and general working-class immiseration. Workers began to organize in large numbers using the tools they still use today. They struck, collectively refusing to work unless employers increased their pay or bargained collectively. And they boycotted, urging other workers and consumers not to patronize non-union businesses.

The legal issue should be clear: strikes and boycotts are, quite literally, combinations in restraint of trade. That is their point. Industrial workers’ only means of improving their lot was to create a cartel, fix prices for labor, and exert economic pressure on anyone who refused to accept their terms. Before modern antitrust, employers often described unions as “labor trusts,” cast their efforts as “civil conspiracy” under state law, and obtained broad judicial injunctions against strikes and boycotts that thwarted organizing. Once the Sherman Act was on the books, employers quickly argued that it, too, prohibited many union activities. They were eventually successful, and at one point the Supreme Court even enabled an aggrieved company to collect treble damages from workers.

Labor pushed for and obtained an exemption from antitrust liability in the Clayton Antitrust Act of 1914, which labor leaders subsequently called “Labor’s Magna Carta.” The Supreme Court soon held that it, too, applied to workers’ efforts. In one such case, Brandeis dissented, joined by Holmes, arguing that the Court was “imposing restraints upon labor which remind one of involuntary servitude,” particularly given the Court’s solicitous treatment of oil and steel trusts under the antitrust acts.

The Court finally switched gears in the early 1940s, interpreting the New Deal labor acts to immunize most labor organizing and collective bargaining from antitrust liability. Yet the pre-New Deal judicial view of strikes as inherently suspect never quite died. For example, the federal courts and the National Labor Relations Board have whittled away workers’ rights to strike to a thin reed, often in direct contravention of statutory law. 

For its part, Congress later amended the labor act to prohibit certain strikes and boycotts that are unusually effective, in part on the theory that those tactics were just too powerful. In a way, the courts and even the NLRB can see or feel the putative coercion inherent in a strike or picket far more vividly than the structural coercion nested in employers’ common law property rights. To pick just one famous example, Justice Stevens once justified statutory restrictions on labor protest that would be unconstitutional as applied to other actors on the grounds that a picket line “calls for an automatic response to a signal, rather than a reasoned response to an idea.”

Plus, the labor antitrust exemption was never complete, and as a result contemporary antitrust directly restricts worker organizing in important ways. For example, businesses have challenged transnational efforts to raise wages for the world’s poorest garment workers as an antitrust problem, and they have challenged unions’ efforts to build support among community allies via private RICO suits that resurrect the old civil conspiracy doctrine.

Contemporary antitrust also prohibits organizing by independent contractors, a category that includes many truck drivers, gig economy workers, and many other vulnerable workers. The posture of antitrust law in such cases is crucial: it doesn’t just fail to protect workers against employer retaliation, it renders their efforts positively unlawful. Under current law, Uber could likely obtain hefty damages from its drivers if they sought to unionize. Thus, what many workers need today isn’t old-fashioned trust-busting, but a rejuvenation of the old-fashioned antitrust exemption. 

More generally, we need a targeted and multi-faceted approach to employer power. Many of the solutions proposed by commentators in this space, such as prohibiting non-poaching agreements and non-competes in the low-wage labor market, are eminently sensible. Alan Krueger and Eric Posner have also suggested revising the DOJ’s horizontal merger guidelines so that a merger could be blocked if it reduced labor market competition. That makes sense. We ought to be talking about breaking up some of today’s largest firms, or turning them into public utilities, as Rahman has suggested. 

But I would worry quite a bit about a broad-based anti-competition mandate in the wrong hands. That authority could be used to thwart all sorts of innovative worker organizing by vulnerable workers, a disproportionate number of whom are non-citizens or people of color.

•••

There is a serious left/right intellectual division here. A realistic effort to combat structural economic power can’t just act “from above,” limiting the power of the largest economic actors. It also needs to act “from below,” enabling workers to protest, organize, and exert countervailing power once again. After all, what would a world of perfect labor market competition actually look like? Workers would underbid each other for jobs, would offer to work longer hours than their coworkers for the same pay, and would be under constant pressure to upgrade their human capital. Labor movements resist such efforts, but of course they face a serious collective action problem: when individual workers have no resources, they have virtually no choice but to take what employers offer them, regardless of any moral commitments to their compatriots.

As the legal realists argued long ago—inspired, in part, by Oliver Wendell Holmes’ famous dissent in a labor injunction case—there is nothing natural about this state of affairs. If the law enables capital to combine into corporations, and thereby to obtain the benefits of scale and centralized decision making, then it ought to enable workers to combine if, as Holmes wrote, “the battle is to be carried on in a fair and equal way.” That is a powerful argument for labor law reform, as discussed above. But it also suggests that extant debates around employer power and corporate power are too humble, and that the basic allocation of rights within firms ought to be on the table.

Take, for example, the unusual U.S. rule of “employment at will,” under which either party can terminate an employment contract at any time and for any reason that is not otherwise unlawful. As a result, in the run of cases, an employer can walk in one day and tell workers that their wages have been lowered, their hours lengthened, their emails monitored, or their rights to file a class action severely restricted—all without legal consequence. Or take the complex of doctrines carving out a zone of managerial rights that labor law simply does not enter. For example, unions generally have no right to bargain or strike over decisions about the basic direction of the enterprise, including investments or the closure of particular sub-units. 

Employer’s common law property rights are a standing constraint on workers’ rights to protest unfair treatment and otherwise to organize, despite the fact that the whole point of our labor law was to restrict those property rights. And of course, non-union workers have virtually no rights to be consulted over anything. In brief, employers decide, and workers obey.

As a result, a firm—by which I mean, here, the legal entity through which managers act as legal employers—can enjoy significant power over workers without being market dominant, or for that matter even especially large. This is hardly an original insight: as Coase himself argued, the essence of the firm is its use of command relationships rather than market contracting. Indeed, the firm itself, where most of us spend most of our waking hours, is a Hobbesian or even an authoritarian place. The proliferation of covenants and arbitration clauses, and the long-running decline of unions and labor’s share of profits, are symptoms of this underlying and legally overdetermined power disparity. 

Antitrust can’t touch this issue. Yes, less concentrated local labor markets and a ban on covenants will give workers marginally more power to advocate for changes—but their employers can basically say “no” and move on. Worker organizing can help, and many labor scholars are today focused on how to reshape our labor law to once again encourage unionization (more on that in a moment). But at the very least, workers ought to be guaranteed individual rights to protest, stronger collective rights to protest, clear rights to get into court, rights to publicly-funded private attorneys, or even protections against termination except when for cause. Without such minimal protections, nobody could reasonably claim that workers have the real freedom to move among employers.

•••

This brings up a final tension: is the Brandeisian vision really up to the task of addressing economic inequality today? Brandeis joined the majority in 1935’s Schechter Poultry, which declared the National Industrial Recovery Act (NIRA) unconstitutional because the “codes of fair competition” developed under that Act delegated excessive authority to the President. The NIRA itself immunized firms and unions from antitrust liability insofar as they developed such codes. Essentially, it gave state sanction to industrial cartels, in an effort to limit ruinous competition and get the economy back on track. After the case, Brandeis told one of Roosevelt’s advisors that “This is the end of this business of centralization.” The irony here is that a growing number of labor scholars—myself included—have come to believe that the best way to encourage worker organizing today lies in legal reforms to encourage sectoral bargaining or standard-setting, perhaps under the auspices of the Department of Labor. But depending on their design, such processes could trigger similar worries about centralization or collusion between unions and firms.

There is more. Our own “trente glorieuses” were founded on a détente between labor and capital, under which unions—by then institutionalized by the New Deal reforms—accepted business’ general right to manage in exchange for regular raises. Brandeis himself died in 1941, before industrial unions and this system really took root, and well before postwar concerns about “big labor” came to fruition. But it seems fairly likely he would not have loved postwar industrial pluralism, which was a far cry from his vision of small government, small business, and autonomous worker organizing.

That system left much to be desired for radical social democrats as well: in addition to its tolerance of massive corporations, our postwar labor movement too often protected only white men, excluding women and people of color, and often promoted family wage ideology. Postwar unions did, however, significantly reduce economic inequality, by raising industrial workers’ wages, pushing for higher minimum wages, restricting executive pay, and generally limiting corporations’ and capital’s power in the political sphere. 

Which is all a way of saying that there is no perfect solution here. Another Brandeis notion—that states can serve as “laboratories of democracy”—may nevertheless provide a guidepost. Today, state-level reforms to encourage or discourage private sector collective bargaining are almost entirely forbidden—“preempted,” in legal terms—by federal labor law. That means that Texas and Alabama can not do as much as they would like to kill unions, but also that New York and California can not do as much to support them. Meanwhile, our weak federal labor law, and the sheer geographic and economic scope of today’s most powerful firms, make it nearly impossible for workers to build sufficient organization to really counter corporations’ economic and political power.

Greater labor law federalism may then be a very promising option. Strong state-level unions would often be possible if states set the rules, and such unions could, over time, build up to the national level. Moreover, since workers are people rather than commodities, they can only participate in organizing and union governance at the local level. Greater labor law federalism could then encourage unions to serve as robust schools of democracy, or sites for working class education and political mobilization.

The common theme here is the tension between concentrated power and democracy. Unchecked corporate power can become cancerous in a democratic polity—unwanted, intrusive, oblivious to the body’s needs, and seemingly unstoppable. The workplace, too often, is authoritarian, a place where we surrender rights to free speech, assembly, and dissent. In both cases, corporations are sovereigns, not citizens, and the remedy lies in democratic control—from above, via the state, to ensure corporate power doesn’t extend too widely, and from below, by workers, to ensure that power doesn’t run too deep.