Abstract
This article combines two relatively new subjects of antitrust scholarly interest: labor market power and corporate governance. In so doing, it speaks to a number of recent debates that have grown up both inside the scholarly antitrust literature and adjacent to it. First, this article interprets the shift in the balance of power within corporations favoring shareholders at the expense of workers, both in economic-theoretical and historical terms. Second, it lays out the role of shareholders and the common ownership channel as a vector for anticompetitive conduct arising between firms, not just within firms, thanks to profit maximization at the portfolio level rather than the firm level. Third, it evaluates the claim that employer market power has increased, relative to other current explanations for labor market trends. Fourth, it ties rising employer power in labor markets to the increasing significance of common ownership. And finally, it contends that antitrust is a suitable policy remedy to the dual problems of anticompetitive common ownership and increased employer power, provided it abandons the consumer welfare standard and instead elevates worker welfare to an equivalent juridical status.
A range of recent studies have demonstrated that the U.S. economy suffers from a macroeconomic problem of market power.
1
Indicators of rising macroeconomic market power include the following: Increasing markups across industries.
2
Declining worker power and the divergence of pay from worker productivity.
3
Anomalously low corporate investment, given the low and declining cost of borrowing and the seeming abundance of profitable opportunities and production factor underutilization.
4
The explosion of shareholder payouts (dividends and stock buybacks) as a corporate use of funds, in place of investment, payments to upstream suppliers, workers’ wages, and taxes.
5
Declining firm entry, labor market dynamism, and initial offerings on public equity markets.
6
Rising horizontal market concentration and the development of vertically integrated and controlled business models in many industries that have the effect of walled gardens.
7
This article focuses on two particular elements of the overall market power story: common ownership by shareholders of overlapping and potentially competing firms, and the increased power of employers in labor markets. The basic story linking these two things to the overall narrative of rising market power is that during the era in which New-Deal-influenced corporate managerialism was dominant in the U.S. economy, from the 1930s to the mid-1970s, shareholders were relatively dispersed and management structure was concentrated. Conversely, workers were, on the whole, more concentrated than they are now, given their greater status and bargaining position within the multistakeholding corporation. Since the revolution in corporate governance that has been playing out since the 1970s, shareholding has become more concentrated and workers more dispersed, shifting power within the corporation away from workers and toward shareholders. This trend has changed management from being an agent that is relatively insulated from shareholders and beholden to workers to being an agent that is beholden to shareholders and insulated from workers.
Concentrated share ownership is one of the vectors for channelizing capital market power to benefit shareholders. Notably, owners of capital have continued to enjoy high returns, even though the long-term decline of risk-free interest rates and corporate costs of borrowing 8 would seem to suggest that what they supply to the market is not scarce. 9 The ability of concentrated shareholders to sway the decisions of management has been crucial to the maintenance of that capital market power. Moreover, one way in which concentrated shareholders can sway management to increase their return on investment is by inducing them to act in anticompetitive ways, benefiting shareholders’ overall portfolios rather than maximizing the value of their stake in any one firm.
This article has three sections. The first analyzes the evolution of firm-level strategies in light of common ownership. The second addresses what employer power in labor markets, one of the ways in which management operates firms to the benefit of shareholders, looks like. The final section highlights particular fact patterns that illustrate both trends and that may be susceptible to an antitrust remedy.
Common Ownership and the “Corporate Governance Channel”
Current corporate governance ideology holds that a well-governed corporation is one which is governed in the sole interest of its shareholders. 10 That contrasts with an alternative in which corporate managers are answerable to a broader set of stakeholders, including other employees of the firm, consumers, suppliers, or regulators. 11 Incentive schemes are set up ostensibly to induce the sole loyalty of managers, by turning them into shareholders themselves through stock options and other equity holdings. The idea is that by owning instruments whose value increases the more profitable the firm, managers will be incentivized to govern the firm in order to achieve high profits above all other considerations.
Policies that have instantiated that preference for corporate governance in the sole interest of shareholders include the lifting of the Securities and Exchange Commission rule prohibiting stock buybacks, 12 the fact that “performance-based pay” is exempt from the cap on the deductibility of CEO compensation from the tax base of the corporate profits tax 13 and financial deregulation that made hostile takeovers financeable and therefore viable. 14 Entire business models of “corporate raiders” and other “activist” investors are premised on presenting shareholder-friendly policy alternatives to larger institutional shareholders who dominate boards of directors, in the hope of amassing sufficient voting power to extract concessions from management that have the effect of increasing the stock price, often at the expense of a firm’s long-term viability.
Economic-Theoretical Background to Profit Maximization and Shareholder Primacy as Normative Values in Corporate Governance
Models of corporate governance tend to treat the question of serving shareholders’ interests as identical with the firm objective of profit maximization. There are two distinct but related elements of the economic-theoretical background for that presumption: the idea that shareholders would agree unanimously with the objective of firm-level profit maximization (the so-called Fisher Separation Theorem) and the notion that an economy consisting of profit-maximizing firms satisfies the Pareto welfare criterion (the “First Welfare Theorem”). Both ideas rely on perfect competition in the markets where firms compete, which eliminates profits in equilibrium. 15
Less formally, profit maximization on the part of firms is assumed, and the question of whether that is the policy management is inclined to follow (whether because that policy benefits shareholders or not) is swept under the rug. Milton Friedman offered a moral case for profit maximization on explicitly shareholder-centric grounds. That case had two components, which Friedman did not clearly distinguish. The first was that profit maximization was what shareholders would want, and as employees of shareholders, management has a duty of loyalty to them. The second was that such a policy would be in society’s broader interest and not solely that of shareholders. 16
Given that background, the incentive problem, if there is one, corresponds to a divergence of interest between shareholders, who would want the firm’s profits to be maximized (and then paid out to them), versus managers, who have some other motive or loyalty to some other stakeholder. 17
Influential models of incentive compatibility in corporate governance then take as their starting point that a contract needs to be designed with the aim of getting management to do what shareholders would want them to (maximize profits) and they typically conclude that such a contract looks a lot like turning management into shareholders themselves, the better that their inherently unobservable behavior on the job conforms to what shareholders would want if they could observe it. 18
Historical Background to Profit-Maximization and Shareholder Primacy as Normative Values in Corporate Governance
In addition to the economic-theoretical background to normative corporate governance policies, it is important to recount the intellectual-historical one: that midcentury managerial corporations of the type that came into existence in the 1920s, when the major industrial firms formed out of the Great Merger Movement of 1895–1903 broke free of Wall Street control, installing a generation of managers who were not solely the creatures of dominant creditors. 19 The New Deal reshaped those firms during the 1930s, by subjecting their managers to public regulation, 20 when the private regulation once ostensibly done by creditors was viewed to have failed in the 1929 crash and subsequent depression. 21 Notwithstanding regulatory policies that constrained the business models managerial corporations were allowed to undertake, the postwar era was a profitable one for them, given the booming economy. But progressive tax policies made shareholders prefer not to have profits paid out to them, and much more stringent antitrust enforcement, starting in 1938, and extending into the 1960s, took horizontal and vertical mergers off the table as means by which to dispense with retained earnings. Meanwhile, empowered workers had to be kept in line with regular wage increases. One alternative to dividends, horizontal and vertical mergers, and wage increases was conglomerate mergers, which should be understood in this context as shareholder payouts in the form of stock in another firm, which was tax-advantaged relative to dividends. The alternative to dividends, stock buybacks, were banned as insider trading under New Deal-era securities regulation.
In the macroeconomic crisis of the 1970s, such firms, be they conglomerates or just managerial corporations more broadly, came to be seen as bloated monstrosities with wage-price spirals built into them—as machines that generated the stagflationary dynamics that bedeviled a Keynesian consensus. Early corporate raiders made claims to incumbent shareholders that the firms in which they were invested were worth more apart than they were together. The summum bonum of scientific management after 1980 became to focus on core competencies, jettisoning extraneous lines of business, and especially breaking implicit or explicit contracts wherein management made long-term commitments to stakeholders other than their shareholders. 22 According to that account, management had a greater interest in feathering their own nests by buying off stakeholders and growing their fiefdoms rather than delivering value for their true masters. If they did not adjust accordingly, then they were justifiably to be replaced by management who would be more faithful servants. 23 That is the intellectual-historical context out of which the shareholder-manager principal-agent framework came to dominate received wisdom about corporate governance. 24
To summarize, the basic assumption at the heart of shareholder-first management is that what serves shareholders best is profit maximization at the firm level, which means cost-cutting 25 and cutthroat competition with rival firms. The idea of common ownership poses a fundamental challenge to that conception: when institutional shareholders who exercise sufficient power in one firm, such that they are in a position to sway managerial decisions, are also equivalently powerful in that firm’s competitor(s), their effective interest is not in firm-level profit maximization, but rather portfolio-level profit maximization. 26 That entails a different set of policies on the part of firm-level managers who are, in this conception, acting as agents of principal shareholders whose holdings span across firms. 27 Maximizing the value of a portfolio of competing firms entails colluding to keep prices high, wages low, production restrained, and so on—which are all the signatures of a classic cartel.
In that sense, shareholder-responsive corporate governance, instituted on grounds of efficiency at both the firm and economy-wide levels, for both theoretical and historical reasons, is in fact inefficient. It diverts resources away from economic production, into the pockets of shareholders, foregoing profitable opportunities at the margin in exchange for inframarginal profits. The evidence has shown that today’s corporations earn high profits and fail to invest, despite low interest rates and seemingly abundant profitable opportunities. All of that bespeaks a competition problem. The common ownership channel is thus a promising vector by which such shareholder-first business models and practices might propagate.
The Place of Common Ownership in Contemporary and Historical Regulation of Corporate Power
Two final points to conclude this section. First, much of the literature on common ownership to date has focused on nationally-oligopolized industries and institutional investors with large stakes of publicly-traded common (voting) stock in many if not all of the firms in those industries. That is an important context, but this article extends that context to cover privately-held firms and the types of institutional investors who tend to own shares in them: venture capital and private equity investors. 28 One important part of the trend of common ownership in the more traditional sense of public equity markets is that it coincides with the decreasing importance of those markets in financing American corporations. The availability of private equity means that companies do not go public until later in their life cycle, 29 when the main purpose is less to raise capital than to reduce the risk of incumbent owners. We can understand that trend toward reduced reliance on public equity markets as part of a suite of corporate finance reforms that concentrate power within corporations (among incumbent managers and inside investors) without impairing the ability of those corporations to raise capital. Another one is the proliferation of dual-class share structures, with insiders retaining total control of voting power even as they sell off their stake in future profits. One reason why public equity investors might be willing to purchase nonvoting shares is that they are reassured that even, perhaps especially, within a dual-class structure, early investors and insiders who retain control will govern in ways beneficial to shareholders as opposed to management or other workers, given prevailing corporate governance norms.
Second, it is important to note that this focus on shareholding, and common ownership specifically, as a vector for anticompetitive business conduct and market structures is not the major departure from antitrust tradition it has been considered to be in the recent literature. The Great Merger Movement of 1895–1903 was a movement wholly driven by the same incentives and a similar institutional basis as today’s common ownership. Merging loose cartels prone to defection and dissolution into unitary entities was a feasible legal option following the weakening of state incorporation laws that had earlier invited challenges from outside shareholders, as well as the nonenforcement of the Sherman Act following U.S. v. Knight. 30 Common ownership—the collaboration of creditors across multiple state-incorporated and (ostensibly) state-delimited firms in an industry, to the point of operating a de facto national monopoly, was the mechanism by which that expansion of corporate power took place. 31
The Clayton Act prohibited interlocking directorates thanks to the experience of the Great Merger Movement, viewed (correctly) at the time as Wall Street’s strategy for protecting creditors from “ruinous competition.” The merger movement is usually considered to have ended in 1903, with the Supreme Court’s ruling in Northern Securities suggesting that federal power might be placed in opposition to such “Morganization.” 32 But the Panic of 1907 convinced the Roosevelt Administration to reverse its antitrust policy and favor consolidation as a financial stabilization policy, submitting both to J. P. Morgan’s resolution to the crisis itself and to his and Wall Street’s larger view of what American corporate capitalism should become. Even now, interlocking directors strongly predict common ownership across pairs of firms. 33
Testing the Clayton Act, the power of Wall Street to direct anticompetitive behavior was at issue as well in the Department of Justice’s 1914 suit against the New Haven Railroad for merging with the Boston and Maine Railroad. 34 Moreover, in the larger intellectual context, economists at that time generally held shareholders, as opposed to managers, to be the source of the economic threat posed by powerful corporations. They conceived managers as scientists or engineers (and thus akin to themselves), whereas creditors were at work to undermine their efforts through cost-cutting and schemes to swindle the public. 35 Economists of that era were skeptical of antitrust and of the Sherman Act, seeing its effect as to inhibit economic and technological development under conditions of increasing returns to scale, which they considered equivalent to an anachronistic and futile attempt to resurrect a lost world of smallholder competition. But they did perceive a threat in shareholder control, especially where that control extended across firms in an industry, which they feared was being used to restrain production and divert its fruits into Wall Street’s pockets. 36
The Clayton Act’s passage, and subsequent enforcement efforts, were a response to the many failures of policy to govern excessively powerful interstate corporations, many of them monopolies. Its authors recognized creditors had been able to stymie previous attempts to control corporate power through coordination at the shareholder level, which is to say, above the level of state incorporation law, thus extending corporate power in the national economy beyond the scope of any one state, or states in general, to regulate. 37 Of course, it turned out that the Clayton Act became incorporated into the Rule-of-Reason regime of antitrust jurisprudence it had been enacted to overturn, until the late 1930s when creditor control of major industrial enterprises had been in retreat, as noted above.
Employer Power and Anticompetitive Conduct in Labor Markets
A great deal of recent evidence now establishes that employers systematically enjoy and exercise market power in labor markets. Firm-level residual labor supply elasticities are well below infinity, 38 workers are increasingly tied to existing employers due to the paucity of outside job offers, 39 and consequently earnings in any one job as a function of tenure are flattening over time. 40 Formal 41 and informal 42 restraints tie workers to particular employers and prevent them from switching jobs, while labor demand and job creation have been low in the recovery from each of the last two recessions. 43 Vertical restraints that bind franchisees and gig economy “independent contractors” should be understood as part of the suite of both indicators of and contributors to the monopsony power of dominant firms and employers. Meanwhile, wages are increasingly firm-specific, resulting in similar workers being paid very differently depending on the firm at which they work. 44 The result of all of that has been an economy-wide decline in the labor share of income. 45
It should be noted that studies interpreting findings of high variation in firm-specific wages for otherwise-similar workers through the lens of a model tend to rationalize those facts as nonetheless at least partly caused by worker rather than firm characteristics, or worker-firm-specific match surpluses, resulting in higher residual supply elasticity estimates once the firm variation component is thereby reduced, though still well below infinity (implying employers nonetheless exercise a good deal of control over pay). 46
All of these findings are what may be expected from an increasingly monopsonized labor market: low hiring, few outside job offers, a weakening job ladder, low wages, low labor shares (at both the macro and micro level), increasing firm-level wage determination (as opposed to market-level wage determination), and perhaps others, in addition to employer concentration in labor markets.
To date, two main interpretations of these facts, other than one based on employer monopsony power, have been advanced: that declining aggregate labor shares represent the reallocation of production toward superstar firms, not market power, 47 and that worsening economic outcomes for workers in fact represent workers’ deteriorating bargaining power, as opposed to employers’ increasing market power. 48 I consider each of these two objections in turn. My basic view is that neither of these objections is consistent with the full set of facts listed above. Employer monopsony power works better as an overarching explanation.
Superstar Firms?
Autor and coauthors propose that the apparent indicators of a macroeconomic increase in market power are better explained by a model of firm heterogeneity and reallocation of production across firms in response to exogenous circumstances that make the economy more competitive, not less. 49 The two specific trends they focus on are rising product market concentration at the national level and the declining labor share of macroeconomic output.
To explain those trends, those authors draw on models of firm heterogeneity from the industrial organization and international trade literatures, that firms differ as to their cost of production, an inherent characteristic of individual firms. Changes to the economic environment cause production to be reallocated across firms, with some inefficient ones shrinking and exiting when output market competition becomes more intense and more efficient ones growing and gaining market share at their expense. 50 These are what Autor and coauthors call “superstar firms.” Those models of firm heterogeneity all differ fundamentally from Cournot oligopoly or oligopsony, in which a greater level of concentration in the market would portend reduced competition.
In models of firm heterogeneity, cost of production, generally due solely to a labor input, is exogenous to the firm. The firm realizes a productivity level from some distribution and then selects its output accordingly, given that cost draw, the overall production technology, and residual demand. This setup requires that labor markets be perfectly competitive, which means that employer market power or any kind of bargaining friction or imperfection in the input market is ruled out by assumption.
The reason “superstar firms” have low labor shares, given that all firms take wages as given in a competitive labor market, is that larger firms can charge higher markups over marginal cost because they face inelastic residual demand curves. The idea that firms with higher market share charge higher markups over their lower marginal cost is an implication of a crucial assumption on customers’ utility functions and hence on the slope of the demand curve: larger firms have more market power because they enjoy lower demand elasticity, by virtue of operating on the inelastic portion of the market demand curve.
There are several problems with the overall interpretation of labor market outcomes that does not depart from perfect competition in input markets. By construction, a model based on that assumption cannot explain the many labor market outcomes referred to at the start of this section, since it is not a model of the labor market, let alone one that admits labor market imperfections. Thus, for example, since production is being reallocated across firms in response to an increase in competition, the model would imply higher rates of labor market mobility—workers moving from relatively unproductive to productive firms. In fact, labor market mobility has declined, and specifically, the rate of workers moving from relatively unproductive to relatively productive firms has declined. 51
Such a model would also predict higher pay for workers overall (even given declining labor shares), since they would be employed in more productive workplaces at which their marginal product is increased, and therefore, thanks to perfect competition, their wages would have to be increased as well. Wages have in fact been stagnant for most workers, and the pay increase associated with switching jobs has, if anything, gone down at the same time as the rate of job switching has also declined.
But the biggest empirical challenge to the superstar firm hypothesis as it relates to labor market outcomes is that wage disparities at the firm level have increased substantially. 52 In fact, the rising firm specificity of wages has been offered as a mechanism justifying, rather than contradicting, the superstar firm story in the first place. The conclusion of the article by Autor and coauthors, “The Fall of the Labor Share and the Rise of Superstar Firms,” itself refers to firm-level wage inequality and the fissured workplace as evidence in favor of the superstar theory. 53 But in a perfectly competitive labor market, workers get paid the same wage no matter which firm they work for, meaning that wage segregation and workplace fissuring operating at the level of the individual firm are inconsistent with that paper’s model of superstar firms and reallocation under increased competition. Firm-level wage determination cannot be reconciled with a model of the labor market that presumes perfect competition.
It is not surprising that the superstar firm story does not explain labor market dynamics, because it is not really an attempt to model the labor market. But then the idea that it offers a macroeconomic interpretation of the declining aggregate labor share (and other evidence that workers are on the whole, worse-off), while at the same time ruling out power dynamics in the labor market by assumption, should therefore be approached with skepticism. Unfortunately, problems explaining labor market outcomes are not the only ones plaguing stories premised on superstar effects.
As stated above, “The Fall of the Labor Share and the Rise of Superstar Firms” was conceived to explain primarily two stylized facts: rising output market concentration and the declining aggregate labor share. In its pre-publication form, the paper posited an exogenous increase in consumer demand elasticity, resulting in lower profitability for inefficient incumbents who could not afford to remain in the market without pricing power to protect them. 54 That version of the model, therefore, predicted declining aggregate markups. Since then, several studies have shown increasing markups in aggregate. 55 Therefore, the paper was revised prior to publication to account for an additional fact beyond rising concentration and declining labor shares: rising aggregate markups. 56 In order to derive that prediction, the model needs an additional element beyond those included in its predecessors.
Models of firm heterogeneity typically have more productive firms charging higher markups than less-productive firms. 57 Therefore, a reallocation of production to more efficient firms will, ceteris paribus, increase markups. But when an entire market becomes more competitive, the average markup declines because all firms’ pricing power declines. In their paper “Market Size, Trade, and Productivity,” Marc Melitz and Gianmarco Ottaviano propose a model in which the net effect of increasing competition is a reduction in the aggregate average markup because the direct effect of competition on overall pricing power of all firms outweighs the reallocation effect concentrating production in high-markup firms.
The opposite is true of the final publication version of “The Fall of the Labor Share and the Rise of Superstar Firms”: the aggregate markup increases in response to increased competition because the effect of reallocation outweighs the effect of a reduction in all firms’ pricing power. The reason why is the latter paper’s assumption that the distribution of firm-level productivity is even more skewed than a Pareto distribution. In that case, the difference in markups is so great between high-markup and low-markup firms that the firm reallocation effect on aggregate markups outweighs the overall competition effect. Thus, the aggregate markup goes up rather than down in response to increased competition. For that reason, the paper by Autor and coauthors achieves an unusual trinity for the empirical industrial organization literature: rising competition, rising concentration, and a rising aggregate markup.
The superstar firm account predicting an increase in the aggregate markup in response to an increase in competition thus relies very heavily on the idea of reallocation from inefficient to efficient firms as the mechanism driving changes in aggregate outcomes. The final problem with this account is that in fact interfirm reallocation does not appear to be a significant response to variation in competitive environments. Specifically, empirical industrial organization has documented that more competitive markets do tend to have more efficient firms. But that is because the firms themselves become more efficient in response to competition (an effect that is not modeled in the paper by Autor and coauthors), rather than a reallocation of production from inefficient to efficient firms. 58 Variation in competition appears to have little effect on reallocation, contrary to what Autor and coauthors assume.
In summary, the account of superstar firms drawn from the empirical industrial organization and trade literature on firm heterogeneity does not suffice as the grand interpretation of macroeconomic outcomes (rising concentration, declining labor share, rising aggregate markup) its authors propose. It is not consistent with the industry- and firm-level studies such models arise out of, nor is it consistent with the wide array of empirical facts to be explained, least of all when it comes to labor markets.
Worker Power versus Employer Power
The other interpretation that has been put forward against the claim of rising employer market power is that employers have not gained power so much as workers have lost it. In the paper “The Declining Worker Power Hypothesis: An Explanation for the Recent Evolution of the American Economy,” Anna Stansbury and Lawrence Summers claim that the story of declining labor share is less the creation of new monopsony “rents,” as a declining firm-specific labor supply elasticity would suggest has taken place but rather a reallocation of existing rents from workers to employers. 59 In other words, employer monopsony power suggests workers are paid less than their marginal product of labor, and the contention that it has increased over time means that the wedge between what workers are paid and their marginal product has grown over time. By contrast, the wage-premium-based theory Stansbury and Summers prefer is that workers are, in general, paid more than their marginal product of labor, and over time, the wage premium that workers are paid has declined in favor of employers.
Some aspects of a wage-premium-based theory are consistent with the facts. It would imply that by laying off incumbent workers and replacing them with lower paid workers who are just as qualified, employers could increase their profits. We know that that indeed happens. 60 Second, Stansbury and Summers point to a declining Non-Accelerating-Inflation Rate of Unemployment (NAIRU; the macroeconomic concept of the minimum unemployment rate consistent with stable wages and inflation) as evidence in favor of a loss worker bargaining power. That is indeed what such a model would predict: a narrowing welfare differential between employed and unemployed workers as the positive relationship between wages and the unemployment rate is steadily weakened. But, contrary to the claim in that paper, it is not the case that rising monopsony power fails to predict a declining NAIRU. In fact, a monopsonized labor market would be characterized by workers agreeing to take whatever jobs are available, even if they pay subcompetitive wages. 61
What is not consistent with both stories is the decline in hiring and in labor market mobility. If workers’ bargaining power has declined at the same time that firm-specific labor supply elasticity is constant, that implies a larger share of the surplus generated by employment matches would be earned by employers rather than workers. In a model of wage posting or of Nash bargaining over wages in a frictional labor market, a reallocation of surplus from worker to employer would cause an increase in job posting in order to satisfy a free entry condition. More jobs have to be created in order to exhaust profitable opportunities for employers or would-be employers. By itself, then, such a model would predict that a declining labor share would correspond to increased hiring and increased labor market churn—workers switching often between low-paying jobs, in which firms increasingly exercise monopsony power by trading quits off against wages. In essence, to use Albert Hirschman’s terminology, the idea of declining worker bargaining power within firms would entail trading off voice in favor of exit. 62
In reality, however, both hiring and worker mobility exhibit secular declines during the period in which Stansbury and Summers point to the decline in worker bargaining power. Again, within the context of a wage-posting model or a model of a search-and-matching process with the wage determined by Nash Bargaining, that fact would indicate a failure of the free entry condition to hold—vacancies that would be profitable to post are not, in fact, posted. That failure of the free entry condition bespeaks rising monopsony power rather than the surplus reallocation those authors propose.
Stepping back, although the authors mark out a distinction between the “declining worker power” and “increasing employer power” accounts of labor market trends, it is hard to make that distinction work in practice. The idea of a declining wage premium above a worker’s marginal product versus an increasing wage penalty below a worker’s marginal product depends on the reality of the marginal product of labor as a coherent, and, crucially, observable labor market concept, either as it pertains to any one worker or to the labor market in aggregate (or anywhere in between). Whether the marginal product of labor does in fact exist, let alone in such a way that it can be measured and distinguished from the wages that are actually paid, thus providing a test between these ostensibly competing theories of how labor markets work, remains in serious doubt. Many aspects of the interpretation of the labor market put forward in this article and by other scholars of monopsony bear close resemblance to the account by Stansbury and Summers. For example, “the increase in shareholder power and shareholder activism has led to pressures on companies to cut labor costs, resulting in wage reductions within firms and the ‘fissuring’ of the workplace as companies increasingly outsource and subcontract labor.” 63 That is, in essence, the narrative offered in this article as well.
It would probably be more fruitful to understand diminishing worker bargaining power and increasing employer monopsony power as both correct, and the relative balance between the two as what matters for determining outcomes and for prescribing policy solutions. In fact, one of the strengths of Stansbury and Summers’s paper is exactly that it centers the same corporate governance channel as is emphasized here—it is not simply that employers are more powerful vis-à-vis workers than they once were, but also that the internal structure of employers has changed, to be more solely responsive to shareholders and less to other stakeholders, including workers.
Where this article takes the argument further, however, is in pointing out that it is not simply that shareholders want to squeeze as much out of the firms they own as they can, at the expense of workers, but, firstly, that they seek to do so at the portfolio or industry level, not simply at the level of the firm, and secondly, the increased size and concentration of shareholders enables them to wield greater influence at the industry or market level than they have been able to since the era of the Great Merger Movement and the Morganization discussed in the previous section, exacerbating the unequal bargaining power between any one employer and his or her workers. That is part of what makes the case for an antitrust response, in contrast with the skeptical stance toward antitrust as a remedy adopted by Stansbury and Summers, as well as similar work.
Antitrust as a Remedy to Common Ownership and Labor Market Power
So far, this article has done two things: explicate the corporate governance channel to the proliferation of anticompetitive conduct and structures across firms and industries and characterize employer power in labor markets as a field in which that anticompetitive conduct and structure is manifest. But what about the linkage between those two things? And Is antitrust a potential remedy?
A simple answer to the first question, identifying the linkage between common ownership and poor labor market outcomes for workers, is that we would not necessarily expect to see it in a way that’s distinct from what we observe from shareholder-first corporate governance. Managers responsible to shareholders would not invest, expand capacity, post job vacancies, or pay higher wages for fear of a hostile takeover bid from activist investors seeking to sway institutional shareholders with decisive voting power. If worker productivity increases, managers who are answerable to shareholders will hand over the marginal profit rather than increase wages or hiring if they can possibly get away with it, which the evidence of labor market monopsony shows that they can. What looks like good corporate governance, then, simply entails worker-unfriendly corporate policy. 64 That is why even though the linkage between common ownership and monopsony power in labor markets is unexplored in the literature relative to the effects of common ownership on consumer-facing outcomes, in some ways it is the more plausible channel for explaining the overall macroeconomic increase in market power. Exploiting an imbalance of bargaining power favoring employers is exactly what managers beholden to shareholders are hired to do under entirely orthodox corporate governance principles.
The reason we might look to antitrust in a broad sense for a remedy is that the employer power fact pattern operates in many instances across and not solely within firms, through the corporate governance channel, because institutional investors are larger and with more overlapping portfolios than had previously been the case. Research in management science shows that shareholders can and do wield that influence, specifically over executives, in such a way that cannot be understood other than as a portfolio-value-maximization strategy. 65
Consider the bank-holding company Softbank, a major shareholder in many companies around the world but specifically in the technology sector. In the last decade, Softbank has typically operated something like a venture capital firm, taking large if not majority positions in private, unprofitable companies and seeking its return when they go public, whether they have achieved profitability by that stage or not. The companies it has funded have used the resulting financial resources to cut prices and expand operations to achieve dominant market shares in particular segments (e.g., ridesharing or short-term office space rentals), at the same time cutting wages and terms of work for a largely contingent workforce and implementing the type of surveillance and algorithmic management techniques that have come to typify the gig economy. 66
In fact, regulators report that founders of startup firms in the technology space are told they have no chance of getting funded unless they control their workforce in this way without legally employing their workers. 67 Gig economy platforms could themselves be understood as inherently anticompetitive, since they control and coordinate markets and allocate customers to workers unilaterally, displacing competition among providers and setting the terms of these transactions to which the platform is putatively a neutral third party. 68 Because the nature of the gig economy is to make use of the independent contractor classification in labor law, it inherently implicates antitrust as the instrument we have to regulate the exercise of power across firm boundaries. 69
But even aside from the gig economy platform business model itself, which might be understood as suppressing intra-brand competition in an antitrust sense, Softbank has acted to suppress interplatform competition among its portfolio firms as well. Softbank portfolio firms have engaged in stock swaps that allocated entire national or multinational ridesharing markets to one company, where there previously had been competition: Uber and Didi Chuxing in China 70 ; Uber and Grab across Southeast Asia. 71 Under circumstances in which each of these platforms is separately unprofitable, the move to consolidate monopolies and suppress competition between them to stabilize creditors’ stakes is quite reminiscent of the Great Merger Movement. It is also undertaken repeatedly in different market segments by private equity firms implementing “rollups,” in which all the firms in a given segment are merged into the same private-equity-controlled entity. That strategy has proven especially successful where, prior to consolidation, portfolio firms exhibited variation in labor share and working conditions, private equity achieves standardization through a race to the bottom. 72
Another example of common ownership’s threat to competition implicating labor markets is the recent acquisition by Liberty Media of IHeartMedia. Liberty Media is a holding company that already held controlling interests in many content properties, particularly related to noninteractive music streaming (e.g., Pandora and SiriusXM satellite radio) as well as live performance (Live Nation/Ticketmaster, itself the product of a controversial vertical merger in 2010, after which performing artists were induced to also use Ticketmaster’s ticketing services for concerts at Live Nation venues.) 73 IHeartMedia owns hundreds of radio stations, as well as its own concert and music festival promotion companies and podcast network. By controlling all of these entities, Liberty Media is now able to reduce royalty rates for streaming music and leverage its control of live performance to extract favorable terms from artists who seek to be promoted on its streaming platforms. 74 If any one of these properties were seeking to maximize profits independently of the others, they might seek an exclusive contract with a performer by paying him or her more, to withhold that performer from the others. That’s what firm-level profit maximization would entail, even if entirely shareholder-responsive. Portfolio-level profit maximization, on the other hand, means that any one of these outlets can secure an artist’s services without paying for an exclusive contract but rather withholding access to the others to enforce one. As in other segments like ridesharing, the “rollup” of outlets by which musicians and other creators used to reach their audiences by one dominant holding company threatens to shut down this formerly vibrant ecosystem by channeling all revenue to its dominant shareholder.
Antitrust is not the only policy tool available to remedy employer power, but it is the one that has historically been successful at regulating the exercise of power across the boundaries of the firm. 75 Institutional investors like Softbank and Liberty Media, which are self-acknowledged active managers, could simply be barred from owning more than one firm in an industry by extending the logic of the existing Clayton Act ban on interlocking directorates. For investors that seek a market portfolio through a rollup, the appropriate remedy is probably an enforceable restriction on the exercise of voting power or the dilution of voting rights as a function of the extent to which a given shareholder also holds competing firms.
Particular enforcement actions, such as a monopsonization claim against a platform or platforms that had agreed to limit competition between themselves, including by effectively enforcing exclusivity on drivers or content creators, are certainly an option. The antitrust case United States v. Apple (2015), 76 regarding Apple’s e-book platform, offers a potentially useful precedent. In that case, Apple orchestrated what the court agreed was a horizontal conspiracy among book publishers seeking an alternative to Amazon’s dominance of the e-book sector. The stock swaps that Softbank has orchestrated among its portfolio ridesharing firms could be understood as equivalent to the hub-and-spoke conspiracy between the publishers and Apple that was at issue in that case, dividing the market into non-overlapping national rideshare monopolies that have been able to reduce wages and working conditions for rideshare drivers.
However, any antitrust case targeting the abuse of power in labor markets faces a formidable obstacle: the consumer welfare standard under which antitrust presently operates. The consumer welfare standard invites the defense that whatever harm to competition might be observed in labor markets (or any other upstream market) reduces costs of production and thus redounds to the benefit of consumers, including through lower prices. In fact, it is exactly through the channel of diminishing workers’ bargaining power that the consumer welfare standard is supposed to operate, as evidenced by a number of antitrust publications that elevated the ability to control and coerce workers as the reason why we would expect adopting such a standard to enhance economic efficiency. 77 Furthermore, Ohio v. American Express invites any would-be gig economy platform defendant to point to pro-competitive effects for consumers as a defense to antitrust claims on the labor side of the platform. The fact that existing antitrust law invites this defense is one of the reasons to expect that the common ownership-monopsony channel for anticompetitive behavior would be a more obvious one for shareholder-responsive management than the common ownership-monopoly channel.
An illustrative example of the challenge the consumer welfare standard poses to any antitrust theory of harm centering labor markets and workers is the ongoing litigation concerning the National Collegiate Athletic Association’s (NCAA) restrictions on remuneration for college athletes, in which both district and circuit courts have affirmed that the NCAA’s restraints are legal insofar as they serve the “procompetitive purpose of preserving amateurism and thus improving consumer choice by maintaining the distinction between college and professional sports.” 78 There are numerous other cases where the welfare of workers has been assumed to trade off with the welfare of consumers, and under the consumer welfare standard, the latter is what antitrust protects.
For example, agreements in which a company that was divesting one of its divisions assured that unit’s purchaser that it would not “hire back” workers from the divested entity for a period following the divestiture were ultimately deemed to be necessary to ensure the realization of merger efficiencies and therefore ancillary restraints to legitimate business conduct. 79 And professional associations prohibiting competitive bidding in their bylaws were presumed to disadvantage consumers, despite the defense that they were necessary to uphold quality standards that benefit consumers. 80 In the first case, employers agreeing not to compete for workers was assumed to be good for consumers. In the latter, workers agreeing not to compete with one another was assumed to be bad for consumers. The sole principle uniting these two judgments is that whatever is bad for workers must be good for consumers and vice versa. 81
All of this signals the difficulty with establishing any kind of antitrust liability in labor markets under the consumer welfare standard, let alone doing so through the shareholder channel thanks to common ownership. If antitrust is to be put to work against employer monopsony power, it should, at the very least, adopt a “worker welfare standard” with equal juridical status to the consumer welfare standard. 82
Conclusion
This article has documented three things: (1) the competitive significance of the shift in corporate governance to favor shareholders when shareholders seek to maximize portfolio-level returns rather than firm-level returns; (2) the imbalance of power between employers and workers as one, highly significant sphere in which shareholder power is manifest; and (3) how the nature of the power imbalance plays out across the border of the firm, necessitating an antitrust response, albeit one that is difficult under the consumer welfare standard. To summarize, power has not simply shifted from workers to employers, but also within firms, toward the large, concentrated shareholders who have greater and greater sway over the management of the firms in their portfolios, relative to other stakeholders. The literature on common ownership has established how that can be anticompetitive, and indeed, common ownership as a vector for harm to competition that is cognizable under antitrust laws has a long tradition, just not recently. Conceptually, there is no reason why these two things cannot be understood as mutually reinforcing and threatening to competition broadly.
Footnotes
Author’s Note
This article is a submission to an Antitrust Bulletin symposium on common ownership and its implications for antitrust policy.
Acknowledgments
Thank you to the editors Einer Elhauge, Sumit Majumdar, and Martin Schmalz and to Anna Stansbury, Jonathan Baker, and Ron Knox for their helpful discussions.