Antitrust and Inequality

I. Introduction

As this symposium issue recognizes, the areas of antitrust law and inequality have sometimes knocked together in recent policy debates. From Thomas Piketty’s data-driven Capital in the Twenty-First Century, ¹ to J.D. Vance’s emotional Hillbilly Elegy, ² to the 2016 presidential election, income and wealth inequality have clearly become centrally important political issues. Even among Harvard Business School alumni, 63% believe that reducing inequality should be a “high” or “very high” priority. ³ Concurrently, though less fervently, antitrust law has entered public discourse as a social ordering problem, as large, consolidated “super-firms” ⁴ have grabbed ominously large market shares, limited consumer choices, and threatened to render local provision of goods and services anachronistic. As disquiet grows over their ubiquity and their displacement of local institutions—and sometimes their treatment of customers—some have looked to antitrust laws to slow this trend.

It is thus unsurprising that inequality and antitrust law should be joined from time to time. Unrest in these areas has brewed for decades, received heightened attention after the global financial crisis of 2008, and exploded into politics recently as populist anger. Both contribute to a growing unease among the nonwealthy—say, those in the bottom 95% of income or wealth—that they are alienated from a richer and more powerful class, and that distant barons somehow control their lives and their fate. Prominent scholars, including Nobel laureates Joseph Stiglitz and Paul Krugman, and the late Sir Anthony Atkinson, longtime advocates of poverty relief, seem to be tapping into a century-old fear of monopoly, ⁵ albeit sometimes casually. Why not? Monopoly or oligopoly rents transfer wealth from consumers to producers, which would seem to naturally lead to an increase in inequality. ⁶

But the linkage between inequality and the rise of this new trend towards industrial concentration is not always so clear. For one thing, as Daniel Crane argues, heterogeneity among consumers and producers render it extremely difficult to determine whether on net, industrial concentration redistributes wealth from poor to rich. ⁷ On the consumer side, the rise of these super-firms that seek to dominate markets for internet search, retail, social media, telecommunications, electronics, and seemingly everything important have, despite their ominously large market shares, incontrovertibly produced enormous consumers surplus. Even as Amazon has driven independent booksellers out of business, it is hard to ignore that the fact that a very, very wide swath of consumers have benefitted from low prices. Without a clean counterfactual of a world without Amazon, Google, Apple, Microsoft, Facebook, and other super-firms, it is difficult to separate out the contributions and the deadweight losses created by their dominance.

For another thing, many of these super-firms arise in industries that still have low barriers to entry, and remain vulnerable to inchoate competition. After all, fears in the 1990s of a Microsoft monopoly have proven to be misplaced. In the two decades since the settlement of the U.S. Department of Justice’s antitrust action against Microsoft, ⁸ other tech giants have emerged to challenge Microsoft, and some have already come and gone. In a 2014 speech, former Google CEO Eric Schmidt said, “someone, somewhere in a garage, is gunning for us. I know because not long ago we were in that garage.” ⁹ Competition in retailing has become ferocious, as Amazon and Walmart started from different places and now find themselves competing in the grocery market. ¹⁰ The logical goal of all of these firms is to become the singular provider of a wide range of goods and life services to billions of consumers worldwide. Tech giants, retailing giants, and even pharmaceutical giants such as CVS and Walgreens seem to be capturing ever-larger shares of a wider variety of goods and services, competing vigorously with each other in overlapping markets, ¹¹ and even branching out into the provision of simple medical services such as vaccinations. ¹² The kind of market power sought by all of these actors would be unprecedented, but competition and unremitting threats of new entrants still seem to render the threat of monopolization remote. ¹³ Or, as Herbert Hovenkamp put it, claims that “low prices today [will be recouped by] monopoly profits later…need to be more than an abstract proposition.” ¹⁴

But while it is premature to lay blame on antitrust law for inequality, it is also imprudent to dismiss it. Antitrust law may, as currently practiced, contribute to inequality in a subtle but important way: by contributing to a shift in the capital-labor ratio of some of the most dominant firms. Antitrust law’s singular focus on efficiency ¹⁵ has, unsurprisingly, helped bring about enormous gains in efficiency. What is less obvious is that it has done so in a way that has created an economy that is more capital-intensive, and less labor-intensive. This raises a thorny normative question about the desirability of a far more efficient economy with far fewer jobs. But if we start with the premise that inequality is too high and needs to be reduced, antitrust law should be part of the conversation.

II. A Consequence of an Efficiency Focus

All kinds of firms and industries themselves have become efficient in ways that were scarcely imaginable just a short time ago. A chicken that took 84 days to grow to five pounds in 1984 now requires only 45 days. ¹⁶ Moore’s Law—that CPU efficiency doubles every eighteen to twenty-four months—has, if interpreted more broadly to include other types of gains in electronic miniaturization, held up improbably well over three decades. ¹⁷ And one could scarcely have imagined just a decade ago that Amazon would be able to deliver almost anything to any address in the Continental U.S. within forty-eight hours for a modest annual fee of $129. Even legal services, which have historically been predicated upon personal relationships, are clearly consolidating into fewer and larger firms, and with a marked shrinkage in the overall demand for lawyers. ¹⁸ The list of efficiencies achieved over the past several decades could go on and on. To be a chain operation today—be it retail, dining, coffee, or service provision—requires not just within-firm cultural uniformity, but also a national or international integration of resources, enabling the shuffling of goods, employees, and other assets.

How have these efficiency gains been achieved? To be sure, technological advances have played a starring role; transaction costs have plummeted. Technological advances have also served to reduce the need to hire people to carry out routine tasks and, as machines have become more adept at learning, increasingly sophisticated tasks. ¹⁹ Legal services and radiology are no longer inherently human-centered tasks of judgment, but potentially machine-centered tasks.

Sprinting ahead lockstep with technological advances have been gains in economies of scale. Technological disruption has produced gains in efficiency, but it has also turbocharged the pursuit of economies of scale, enabling firms to marshal vast networks of resources, and deploy them ever more efficiently and, by implication, at greater scale. Big-box stores like Walmart and Costco are very, very large buyers, purchase at volume, and manage their distribution and sale with precision and efficiency, aided by the ability to inventory their goods on a minute-to-minute basis. One can walk into virtually any one of the more than 24,000 Starbucks stores throughout the world ²⁰ and be able to order roughly the same coffee as one would in Seattle. Even livestock farming has consolidated around economies of scale: in 1969, only 7% of all American-grown pigs were raised on farms of 1,000 or more pigs; by 2012, that figure was 96%. ²¹

This increasingly common model of profit-making—using technology to handle and sell large quantities at low profit margins to gain and maintain market share—may or may not yield monopolies. This essay does not address this prospect, but instead focuses on another important industrial implication of this model: that firms are achieving these efficiencies by substituting capital for labor. Producing at large scales requires the reduction of variable costs, which can only be achieved by an enlargement of fixed costs. This model of profit-making requires larger amounts of capital. It has also become obvious that it implies reductions in the most significant variable cost, labor, aided by the technological advances in automation. Reducing the capital-labor ratio, and therefore employment, would appear to have the first-order effect of increasing inequality, though this seems to be an open question among economists. ²²

Laissez faire antitrust policy is not obviously, at this point, increasing inequality by allowing monopolists to amass wealth by gouging consumers. But laissez faire antitrust policy may be contributing to inequality by incentivizing the substitution of capital for labor, changing the capital-labor ratio of dominant firms and possibly the economy as a whole. To be sure, there are certainly other policies that contribute to a change in the capital-labor ratio, and perhaps much more so. But antitrust law offers one view into this trend and an opportunity to consider the implications of a changing economy.

Economist David Autor and various coauthors have written a series of working papers reporting their examination of the causes of what is now recognized as decades-long decline in labor shares in the U.S. ²³ The phenomenon is not confined to the U.S. ²⁴ This violates one of the most time-honored “stylized facts” of economic growth: that the capital-to-labor ratio remains constant. ²⁵ For decades, economists have operated from the assumption that if capital became more productive, its price would rise relative to labor, so that firms would acquire more labor, bidding up the price of labor and restoring a natural equilibrium. Thomas Piketty’s Capital has made a similar, and similarly global argument, that returns on private capital are exceeding economic growth, leading to increasing wealth inequality. ²⁶ Piketty, too, argues that the capital-to-labor ratio has been increasing. ²⁷

III. The Shifting Capital-Labor Ratio

Why has the Earth moved under economic theory? It could be that international trade has driven down domestic wages as jobs are moved offshore to cheaper labor markets. ²⁸ It could be that the products of intellectual property have become more important and have driven down labor shares. ²⁹ It could be that technology has made capital cheaper (relative to productivity, or a better value), and that has caused firms to shift away from labor and towards the relatively more productive and/or cheaper capital. ³⁰ What Autor et al. suggest is that dominant super-firms such as (but not limited to) Google, Amazon, Microsoft, Apple, and Facebook have become much more efficient than their within-industry peers by taking advantage of this change in capital by obtaining more of it, and substituting it away from their variable costs like labor. ³¹ Super-firms use this cost-advantage, which requires large economies of scale, to grab larger market shares. Those unable to seize this advantage lose market share to super-firms, and the result is a more concentrated industry. ³² Google and Facebook serve as prime examples, ³³ but the same trends were identified in six of the major sectors studied by Autor et al.: manufacturing, finance, services, utilities and transportation, retail trade, and wholesale trade. ³⁴ All of the six have experienced increased concentration. ³⁵ Four of the six—excluding finance and wholesale trade—experienced significant changes in their payroll-to-sales ratio, ³⁶ a proxy for the inverse of the capital-to-labor ratio.

To be sure, many, many social, economic, and technological factors have contributed to a change in the capital-labor ratio, and a fixation on efficiency in antitrust law is at most only one of several. Technological change, in particular, might dominate other factors. But technological change is not completely exogenous. Technological change could be capital-augmenting—something that makes capital more productive or cheaper—or could be labor-augmenting, and could incentivize the hiring of more workers. ³⁷ Importantly, technological change could also be path-dependent, as capital-augmenting technology could make capital more efficient, and therefore beget more capital-augmenting technology (so too for labor-augmenting technology). So it is quite possible that fairly small policy preferences—such as antitrust law and policy—could result in fairly significant economic changes, and could significantly alter capital-labor ratios.

The second step in this analysis is also uncertain: Does an increase in the capital-labor ratio lead to income or wealth inequality? A first-order economic analysis would certainly suggest so: Fewer jobs would push wages down, and more capital would certainly suggest higher returns to capital. But as with many other economic matters that seem obvious, it’s complicated. ³⁸ Higher returns to capital might be reinvested so as to create more, and possibly better jobs. Technology could also make workers more productive, restoring an equilibrium temporarily disturbed by some spurt in technology. ³⁹ Or, even if the Autor studies ultimately prove to be the last or nearly-last word, there is the response, “So what?” Isn’t the competitive behavior of super-firms what the goal of antitrust law was supposed to be? ⁴⁰ Perhaps. But if one accepts that inequality is a social ordering problem in need of redress, then some discussion of the normative underpinnings of antitrust law would seem to be called for, even overdue. If even just a prima facie case exists that higher capital-labor ratios are a fundamental shift and not a temporary phenomenon, then at the very least some further research and discussion is needed to examine the role of antitrust law in helping to bring about a potentially structural change to the economy. A richer economic analysis would inform antitrust lawmakers, who now need to take into account a much broader range of economic considerations, and not just consumers surplus.

IV. Whither, Antitrust Law?

Given all that, can it still seem incongruous for antitrust law be reshaped to address inequality and play a role in allocating wealth? It may seem simply too much to put on antitrust law to expect it to reverse a trend as fundamental as a shift in capital-labor ratio. For one thing, it is impossible to ignore the consumer benefits that have flowed from some super-firms such as Amazon and Walmart, or the information benefits from tech giants such as Google. For another thing, it would seem odd for antitrust law to be contorted to remedy something so out of its wheelhouse as economic inequality, which is still somewhat removed from the problem of the industrial titan monopolist that the original Sherman Act sought to fix. ⁴¹

And then, finally, there would seem to be a number of other areas of law that are more explicitly and more directly oriented towards the allocation of income and wealth. Tax law leaps to mind, especially insofar as it affects capital investment. ⁴² Piketty also points out, as others have pointed out, ⁴³ that at least in the United States, a steep personal income tax cut for high-income individuals has helped produced a class of “super-managers,” high-income earners who earn nonsensically large executive salaries, unmoored from performance. ⁴⁴ If, as seems likely, a regressive personal income tax structure contributes to inequality, then personal income taxation would be ripe for reform. So too, estate tax reform. And if aggregate employment is truly a root problem underlying inequality, then employment law may be an area of focus, even if it runs counter to antitrust law. For a problem as complex as inequality, the list of possibilities is long.

But even if antitrust law does not get a starring role in reducing inequality, it should not get a pass. Laissez faire antitrust law could be an important factor in the industrial drive to consolidate and concentrate. As Eleanor Fox put it, “[t]he operational goal…is to let business be free of antitrust unless its acts will decrease aggregate consumer surplus.” ⁴⁵ Price benefits are, in this environment, easy to demonstrate, so this a very significant safe harbor, and may very well play an important role in changing the capital-labor ratio.

Certainly, some expansion beyond the most simplistic notions for efficiency are called for in an antitrust jurisprudence and enforcement policy. The late Sir Anthony Atkinson has argued for including a distributional component in antitrust cases and adjudications, though he does not further elucidate. ⁴⁶ Antitrust scholarship has seen numerous calls in the past for something less reductionist than a straight economic efficiency test, ⁴⁷ like preserving competition for its own sake. ⁴⁸ The notion that courts and enforcement authorities should expand their notions of efficiency is hardly a new one. Moreover, at some level, inequality is allocatively inefficient, ⁴⁹ so an expansion of efficiency to include additional considerations that bear on inequality could be within the bailiwick of many antitrust scholars.

I leave for future research and scholarship a detailed discussion of what such a broadening of antitrust law would look like. However, a few general thoughts are in order. First, much of the angst over both antitrust law and inequality center upon the increasing prominence of capital and the declining importance of labor. With that curious parallelism in mind, it is surely worth thinking about efficiency more broadly, taking into account the economic effects of job losses. Certainly, jobs are a central consideration in all manner of regulation, even if the costs of unemployment are left unquantified. ⁵⁰ It is not as if just considering employment effects in horizontal merger analysis would be an impossibly foreign task to antitrust lawyers and economists. The Department of Justice and Federal Trade Commission Horizontal Merger Guidelines ⁵¹ are silent on the subject of jobs in considering horizontal mergers. Not once do the words “job,” “employment,” “employee,” or “worker” appear. The Guidelines provide several examples of merger-specific efficiencies that would be recognized as justifying a decrease in competition: the creation of new products, ⁵² combination of innovation efforts, ⁵³ the salvage of failing firms, ⁵⁴ and the combination of complementary assets, ⁵⁵ to mention just a few. At best, the Guidelines express complete indifference as to whether a horizontal merger would eliminate jobs or not. At worst, the Guidelines, in demanding greater and more verifiable efficiencies to justify mergers, push competitors towards job-reducing mergers. Because evidence of efficiencies is sometimes hard to come by, the Guidelines even seem to privilege horizontal mergers that could eliminate jobs. They cite with approval “efficiencies from shifting production among facilities formerly owned separately, which enable the merging firms to reduce the incremental cost of production, are more likely to be susceptible to verification and are less likely to result from anticompetitive reductions in output.” ⁵⁶ None of this is necessary in a detailed discussion on merger-specific efficiency. Sources of efficiency that do not eliminate jobs can be distinguished from those that do, and may warrant more deferential treatment.

As well, it may be worthwhile to think about capital and question the efficiency of ever-larger capital. As I have argued elsewhere, larger capital means that the owners of that capital have more at stake, which means larger rent-preserving activities, which are quite likely to be inefficient. ⁵⁷ In my reading of Mancur Olson’s Rise and Decline of Nations, the underlying premise behind Olson’s hypothesized one-way ratchet of increasing unemployment, stagflation, and economic decline of nations stems from the impulse of capitalists to protect their capital, ultimately to the detriment of a broader society. ⁵⁸ Perhaps antitrust lawmaking (nor many other forms of lawmaking) need not be so solicitous of entrepreneurs wishing to accumulate or protect capital. In Verizon Communications v. Trinko, ⁵⁹ Justice Scalia wrote that monopoly was “an important element of free enterprise” and that the opportunity to charge monopoly prices—at least for a while—is what attracts “business acumen” in the first place. ⁶⁰ Monopoly rents are exactly what the Sherman Act was intended to prevent, so reading that out of the statute requires quite a bit of chutzpah, even for Justice Scalia. Even with the efficiency framework, it seems eminently reasonable to repudiate the inclusion of private monopoly rents as a necessary incentive for the accumulation of capital. That would appear to be well within a reasonable scope of antitrust law.

Finally, a focus on inequality, and a realization that laissez faire antitrust law may be contributing to an increase in the capital-labor ratio, may be an additional argument for revisiting the efficiency hegemony in antitrust law. There is, after all, nothing more rational about the simplistic notion of efficiency used in antitrust law than other criteria. ⁶¹ I thus follow others in noting that as a matter of legislative history, a defensible interpretation of the Sherman Act would not exclude consideration of other goals that might expand notions of efficiency, or even push beyond efficiency into political goals. ⁶² Indeed, the legislative history of the Sherman Act is rife with statements expressing concern about the concentration of power in the hands of the few, and that monopolists were essentially “extorting” helpless individuals. ⁶³ While the world has moved on from the oil and railroad barons that seemingly choked commerce and politics a century ago, the dangers of modern-day super-firms accumulating too much political power is not such a remote possibility. The strategy of super-firms is not to directly monopolize a single market, but to expand into connected markets, sometimes loosely connected markets. The goal seems to be to become indispensable providers of many things, raising the specter of power concentration. This is especially true with super-firms such as Facebook, Amazon, and Google parent Alphabet that have information and news businesses that constitute a significant fraction of their value. And this is true even if consumer prices remain low. At the root of the economic might of super-firms is their capital, which, coupled with low consumer prices, winds up being the barriers to entry.

At bottom, it is a daunting prospect for antitrust law to move away from a simplistic notion of efficiency. How are courts and enforcement authorities supposed to work through the multiple causal links from industrial practices to broad societal trends? The answer is: quite carefully. That is not such a hot potato if one remembers that a number of rebellious antitrust scholars have been straining for decades against the notion that efficiency should exclude all other considerations in antitrust. ⁶⁴ How exactly that is carried out may have to be determined on an ad hoc basis, with lessons to be learned along the way. The problem of economic inequality is such that it is also infeasible to ignore the contribution of antitrust law to inequality, and its role in altering capital-labor ratios. So, I say, in we go.