To those who are worried about the state of contemporary American politics—those who are concerned about the historically high levels of polarization between the two main political parties, who despair of the disappearance of anything that could be called common ground, who bristle at the apparent unwillingness of any occupant of national, state, or local office to recognize the common sense or basic human decency of any proposal coming from the opposite side of the aisle—we offer you this single harmonious word of relief: antitrust.
A vocal concern for the power held by some of the United States’ most dominant companies—especially tech giants such as Facebook, Amazon, and Google—may be the only shared material among the talking points of President Donald Trump and the Democrats vying to run against him in 2020. Trump has asserted that the US should follow the European Union’s lead in handing down large fines to big tech companies for antitrust violations, and during his presidential campaign, he charged that Amazon has a “huge antitrust problem.” A number of prominent Democrats, including Bernie Sanders and Elizabeth Warren, are on the same page, having suggested that many such companies may need to be broken up. In July, the Department of Justice (DOJ) announced that it was “reviewing whether and how market-leading online platforms have achieved market power and are engaging in practices that have reduced competition, stifled innovation, or otherwise harmed consumers.”
The same month, Facebook acknowledged it was the subject of a Federal Trade Commission (FTC) investigation, which reportedly probed questions such as whether Facebook strategically acquired nascent competitors before they could develop into greater threats.
And in September, attorneys general for 50 US states and territories announced an investigation into whether Google’s dominance in online search and search advertising has created anticompetitive conditions—what Louisiana attorney general Jeff Landry, a Republican, called “an absolutely existential threat to our virtual marketplace.” Google took in nearly 75 percent of all US search-ad revenue in 2018, says data-research company eMarketer. On average, more than nine out of every 10 online searches are done through Google or its YouTube subsidiary, according to some estimates.
Concerns about competition are not unique to the tech industry. Aggregate levels of US industrial concentration—or how market share is divided among manufacturing companies—began to increase in the early 1980s after decades of relatively little change, according to research by Chicago Booth’s Sam Peltzman. The trend continued into the 21st century. Between 1987 and 2007, average concentration—as measured by the Herfindahl-Hirschman index, a commonly used gauge of market concentration—within the 386 industries included in his analysis increased by 32 percent.
If this trend toward more-concentrated industries has been accompanied by a small number of companies expanding their market power as a result of diminished competitive pressures, the effects could be momentous. In fact, some research suggests the exercise of market power could be responsible for everything from higher prices to reduced investment to the steadily diminishing share of the US economy that’s enjoyed by the labor force.
But policy makers seeking to address this issue should first consider a few critical questions. Are markets in fact becoming less competitive? If so, what are the best ways to address corporate giants? Would more-aggressive antitrust enforcement—or even breaking up, say, Facebook—be a real cure for growing market power and produce more-balanced, competitive markets? Granted, there’s little consensus about even the most fundamental of these questions right now—and a compelling need for answers.
Is market power growing?
Between the 1960s and the ’80s, merger policy in the US underwent dramatic changes, transforming, in the words of the late Robert Pitofsky of Georgetown, a former chairman of the FTC, from “the most stringent antimerger policy in the world” to “an extremely lenient merger policy, regularly allowing billion-dollar mergers to go through without government challenge, even when they involved direct competitors.”
In 1982, the DOJ introduced its Horizontal Merger Guidelines, reflecting a shift in the US’s approach to horizontal mergers, or mergers of competitors. Whereas before, the courts had presumed such mergers to be illegal, following the release of the guidelines—which defined a specific maximum level of permissible market concentration resulting from a merger—courts now asked whether mergers would harm consumers. According to Peltzman, this shift proved to be an inflection point in the growth of concentration, at least in manufacturing. Whereas average industrial concentration levels in 1982 looked largely similar to those in 1963, from 1982 to 2002, average concentration among manufacturers increased by 24 percent.
There are other potential explanations for increased concentration. MIT’s David Autor and John Van Reenen, with University of Zurich’s David Dorn, Lawrence F. Katz of Harvard, and Northwestern postdoctoral scholar Christina Patterson—who will join the Chicago Booth faculty in 2020—point to the rise of “superstar firms” as a contributing factor, suggesting that “markets have changed such that firms with superior quality, lower costs, or greater innovation reap disproportionate rewards relative to prior eras.” But regardless of its source, concentration growth could indicate a problem if it reflects that a few dominant companies in an industry are using their influence to keep down rivals.
There is a distinction to be made between high market concentration—in which a relatively small number of companies enjoy a large share of the entire market’s sales—and the presence of market power. Economists define market power as a company’s ability to set prices higher than it would set them under competitive conditions. Such a failure of competition could also mean reduced output, lower product quality, reduced pressure to innovate, and lower wages for workers, among other negative outcomes.
However, if a company or small group of companies controls a large portion of market share, that doesn’t necessarily mean it wields market power. In fact, it could mean just the opposite.
“We know entry has been going down over time, by many measures, and profitability of the incumbents, especially the largest firms, has been going up. All this is consistent with some market power.”
“Concentration isn’t a good barometer of the extent of competition in the market,” says Chicago Booth’s Chad Syverson. “It’s not just a noisy barometer; we don’t even know what direction the needle is pointing. There are cases where, clearly, things happen in a market to make it more competitive, and concentration goes up.”
Consider a market in which customers face high switching costs, or barriers from moving from one seller to another. The mobile-phone market prior to the advent of number portability, or the right to take your phone number with you when switching carriers, was a good example: the inconvenience of transitioning to a new phone number imposed a high cost on those who wanted to leave their provider for a new one. If switching costs come down, the market becomes more competitive, but at the same time, many inferior suppliers will lose market share or go out of business as their customers abandon them for better options, resulting in higher concentration.
Chicago Booth’s Luigi Zingales says that even in concentrated industries, incumbents may need to behave competitively if they fear that new entrants to the market could overtake them. A company’s share of the market may be large, but its ability to manipulate the market could remain small.
So is market power growing along with concentration? One standard measure economists use to study this question is the markup of prices over marginal costs. Marginal cost refers to the cost a company faces for producing one additional unit of its product: if a company produces 500 shirts per week, the marginal cost is the added expense of producing the 501st shirt. That cost would presumably include money for more thread, buttons, and elastic, but also potentially for additional wages, equipment, or space if the company is already operating at capacity.
According to traditional economic theory, in a perfectly competitive market, the price of a good should equal its marginal cost. The gap between price and marginal cost is known as the markup; the higher the markup, the less competition is doing to keep prices at a more efficient level.
Some research in recent years has found evidence that markups have indeed gone up over time. Jan De Loecker of KU Leuven, Jan Eeckhout of Pompeu Fabra University, and Harvard’s Gabriel Unger looked at 60 years of data from publicly traded US companies and find that after decades of stability, markups have experienced two periods of pronounced growth since 1980. The researchers’ findings suggest that average markups hovered around 1.2–1.3 (that is, prices were 20–30 percent greater than marginal cost) from 1955 to 1980, but rose to 1.61 (prices were 61 percent greater than marginal cost) by 2016. Such growth could represent a substantial increase in market power.
De Loecker, Eeckhout, and Unger also find that this trend in markups isn’t visible across all companies—in fact, for the median company in their analysis, markups have stayed largely stable. Rather, the overall average has been pushed higher by companies in the top half of the markup distribution, and especially by those in the top 10 percent, which have seen steep increases in markup levels.
About a quarter of the increase in markups the researchers document comes from overhead expenses—many of which are not factored into marginal cost—having grown over time, they say. They give the example of a tech company that spends heavily to develop a new piece of software, which can then be produced in large quantities at low marginal cost. In this case, the marginal cost doesn’t reflect the huge expense of creating the product in the first place. But “while overhead costs have increased, markups have increased even more and firms charge an excess markup that more than compensates for overhead,” they write.
However, markups are difficult to observe. “There is an issue that price is kind of easy to measure, but marginal cost is a really hard thing to measure,” Syverson says.
De Loecker, Eeckhout, and Unger estimated markups in part by using an accounting measure, cost of goods sold (COGS), that includes all of a company’s costs that are directly traceable to a unit of output. However, University of Chicago PhD candidate James Traina argues that COGS doesn’t include important marketing and management expenses that have grown over time. Factoring those expenses into companies’ costs, “I find that public firm markups increased only modestly since the 1980s,” Traina wrote in a post for ProMarket, a blog published by Chicago Booth’s Stigler Center for the Study of the Economy and the State. “Moreover, this increase is within historical variation—measured markups have increased from 1980–2010 as much as they have decreased from 1950–1980.”
How to measure sales markups
Some research suggests that markups in the price of goods and services have increased only modestly since the 1980s, and that the markup measures should take into account marketing and management costs, which have grown over the same period.
Markups are not the only way to measure market power, however. “Profits are a better indicator (of market power) than markups,” Zingales says, adding that it’s problematic if a market has highly profitable companies but few new entrants. “We know entry has been going down over time, by many measures, and profitability of the incumbents, especially the largest firms, has been going up. All this is consistent with some market power.”
This, too, invites debate. The growth, or lack thereof, of profits is disputed territory in the economics literature. In research he started as a PhD student at Chicago Booth, London Business School’s Simcha Barkai examined what portion of US economic output has gone to labor, capital investment—such as the facilities, machinery, and technology used for production—and profit over time. He finds that labor’s share of the economy has gone down, a prominent trend in the economics literature that was first documented in 2013 by Chicago Booth’s Loukas Karabarbounis (now at the University of Minnesota) and Brent Neiman. On average, Barkai writes, workers’ share of sales dropped by 10 percent between 1997 and 2012.
But he also finds that the capital share—traditionally assumed to rise when the labor share falls—has also declined. Instead, what has risen is the profit share: according to Barkai’s research, profits increased more than sixfold from 1984 to 2014, from just over 2 percent of gross value added to the economy to almost 16 percent.
Barkai also used census data to identify the share of sales in each industry accounted for by that industry’s four, eight, 20, and 50 largest companies, and used those shares as measures of industry concentration. He finds that industries that had experienced the greatest growth in concentration also had the greatest decline in labor’s share of sales.
“A decline in the demand for labor inputs (which results in a decline in the labor share) and a simultaneous decline in demand for capital inputs (which results in under-investment) are distinctive traits of declining competition,” he writes. Barkai concludes that declines in the capital and labor shares of income accommodate dramatic growth in profits.
But just as markups are difficult to measure, so is the capital share of the economy, a point emphasized in 2018 research by Karabarbounis and Neiman. If growth in the capital share is underestimated, growth in profits will be overestimated.
Incomplete measure of profits
A pair of researchers challenge a common interpretation that subtracting labor and capital costs from revenues paints a clear picture of the surging share of US business income going to profits since the 1980s . . .
. . . and they point to the virtually inverse movement of the real interest rate, which affects companies’ capital costs and suggests that profits, by the above calculation, are not rising so much as returning to pre-1980 levels, which were higher.
Karabarbounis and Neiman, 2019
Karabarbounis and Neiman note that the standard way many economists measure capital costs is incomplete, as it does not account for risk premia, or how much investors have made by taking risks rather than simply holding Treasury bonds. Changes in risk premia may mean that the financing costs faced by companies do not move in lockstep with the interest rate on US Treasuries. If financing costs have not declined all that much, it would imply that profits have not risen all that much, and would call into question whether companies are really profiting from a rise in market power, at the expense of consumers and workers.
Karabarbounis and Neiman argue that the large growth in profits since the 1980s found by some economists, including Barkai, follows from the assumption that company borrowing costs fell along with US Treasury rates. This is a common assumption in economics research, but if applied to earlier decades, it also implies that profits fell dramatically between 1960 and 1980, when US Treasury rates were increasing. Karabarbounis and Neiman write, “One must acknowledge that the same methodology driving inference about rising profit shares since 1980 reveals that profit share levels in the 1960s and 1970s generally exceeded the levels reached today.”
In that case, is rising market power really pushing up profits, at the expense of workers and consumers? Or, once capital costs are calculated differently, are profits rising less than suggested by these recent studies? Karabarbounis and Neiman lean toward the latter conclusion.
Why would market power be expanding?
But let’s assume that market power is growing broadly. If that were true, what might be causing it? Research suggests a number of policies that may be helping it along—and not all of them fall within the traditional bounds of antitrust.
For example, research from Princeton’s Ernest Liu and Atif Mian and Chicago Booth’s Amir Sufi finds that the low interest rates experienced in the US and many other developed economies since the 2008–09 financial crisis may have contributed to declining competition.
Although low interest rates have traditionally been assumed to encourage business spending, the researchers argue that as rates fall, bigger companies can use them to make bigger investments—in new equipment, for example—than their smaller competitors. These more-significant investments carry bigger productivity payoffs, increasing the competitive gap between the big companies and their rivals. If rates are low enough, eventually both big and small companies will lose the incentive to invest, as the small companies will fall hopelessly behind in market share, relieving some of the competitive pressure on their bigger rivals.
Policy around intellectual property (IP) may also be a contributor to declining competition. Research from Stanford’s Mordecai Kurz suggests that rising monopoly power has accompanied the information-technology revolution, and has been protected by the US’s system of patents and other IP rights.
Kurz documents the growth of surplus wealth—the difference between a company’s total wealth and the capital it employs—generated by publicly traded US companies and finds that greater surplus wealth is associated with companies that have been most transformed by IT innovations. He also finds that the rate of surplus-wealth generation over the years corresponds with various phases of the IT revolution, with more wealth generated during particularly innovative periods of the IT era. These findings, he writes, result from a public-policy regime oriented toward encouraging innovation with the promise of monopoly power.
“To encourage innovations, our laws protect patents and intellectual property rights, granting innovators monopoly power over the results of their innovations,” Kurz writes. “But, once an initial monopoly is established, advantages of first mover together with a mix of updated patents, intellectual property rights and trade secrets, make it very hard for potential competitors to enter the market.”
Some explanations for growing market power do come down to matters of antitrust policy. Part of the evolution of US merger policy in the latter half of the 20th century was the introduction of the 1976 Hart-Scott-Rodino Antitrust Improvements Act, which set up requirements for premerger notifications. The act, writes Chicago Booth’s Thomas Wollmann, effectively meant that for any merger in which the target company was worth at least $10 million, the merging companies needed to notify the government of their plan to merge in advance, so that the US’s two antitrust authorities—the DOJ and the FTC—could consider its likely effects on competition. Deals with the potential to impinge upon competition could trigger an investigation; investigations that indicated some likelihood of competitive harm could lead to attempts to block the merger.
In 2001, the act was amended to bump the prenotification threshold to $50 million. Wollmann finds that following the amendment, premerger notifications dropped by 70 percent. Investigations of deals between the old $10 million threshold and the new $50 million mark dropped from about 150 per year to close to zero—now exempt from notification requirements, they are almost never investigated. But that doesn’t mean they don’t pose any competitive threat.
“Notably, 32 percent of all HSR-related investigations prior to the amendment target deals valued at less than $50 million, rejecting the notion that these newly-exempt deals are unlikely to be anticompetitive,” Wollmann writes.
Moreover, firms responded to decreased enforcement with even more horizontal mergers. Wollmann’s findings indicate that many firms—knowing that they were less likely to face antitrust scrutiny after the amendment—became more likely to propose acquisitions of their competitors.
The mergers may be relatively small, but they are large in aggregate: Wollmann finds that over about a 15-year period, transactions exempt from premerger notification reporting consolidated over $400 billion in US output. Over the 10-year period following its passage, “the threshold increase induces 3,240 competition-reducing business combinations,” a phenomenon he calls stealth consolidation.
Furthermore, although growing concentration is not the same as growing market power, some research suggests that they do share an association. Rice University’s Gustavo Grullon, York University’s Yelena Larkin, and University of Geneva’s Roni Michaely find that companies in more-concentrated industries tend to enjoy greater profits—and that these are driven primarily by higher markups. They also find that horizontal mergers in more-concentrated industries elicit a stronger market reaction than those in less-concentrated industries, and that shareholders of companies in more-concentrated industries enjoy higher-than-average returns. The researchers argue that both of these findings reflect the market’s recognition that companies in highly concentrated industries tend to wield profit-enhancing market power.
The trouble with big tech
Many antitrust discussions today quickly turn to the technology industry, and for good reason. The questions of whether market power is becoming more pervasive and, if so, how policy makers should confront it, are complicated by the growth of digital platforms, for which the economic context is different than that for traditional markets. Traditional policy levers might well be sufficient to fix competition in legacy industries, Zingales says, but digital platforms require “more targeted intervention.”
That’s in part because digital industries such as social media and online search are particularly prone to network externalities—that is, the more people use them, the more valuable their services become. These externalities naturally fuel the winner-take-all (or winner-take-most) outcomes that contribute to industry concentration: the more people who use Google to search the web, the more data Google has to help its algorithm generate the most relevant search results, making its search services even more attractive to future customers. Similarly, when many of your friends post their life updates on Facebook, it becomes more appealing for you to do the same—and Facebook’s popularity is self-perpetuating.
Self-perpetuating market power
Facebook’s and Google’s large market shares are driven in part by network externalities—the notion that the more people use them, the more valuable their services become.
StatCounter Global Stats
“When there are strong network effects, you can get really concentrated markets,” Syverson says. “But the good kind of concentration, where the market is concentrated because everyone likes the best company or the best companies, doesn’t have to exist in a network-goods market. Network-goods markets have a tendency to stay concentrated even if the market leader is inferior to alternatives out there.”
Zingales points out that the network effects at work in the social media industry have a historical analog. In the early days of the telephone industry, customers signed on for service with a specific network, and could only call other customers of that network. Competitive advantages naturally accrued to the biggest networks. The solution, imposed by regulators, was forced interoperability: the networks all had to work together, allowing customers to call each other regardless of which network they used.
Zingales advocates a similar regulatory approach to social media networks. “How do we solve the problem (of network externalities)?” he asks. “We force interoperability, so that I can access Facebook through Pinterest, and vice versa. That will, in my view, restore competition to the market.”
However, network effects are not the only competitive idiosyncrasy of digital markets. In a policy brief reporting on the findings of the Stigler Center Committee on Digital Platforms—consisting of 30 academics and policy makers, who spent more than a year studying how digital platforms affect not only markets but also the media, personal privacy, and political systems—Zingales and Stigler Center fellow Filippo Maria Lancieri identified a number of other concerns that encourage monopolization:
i. strong economies of scale and scope (companies are encouraged to extend a product to more people, or to develop new features);
ii. marginal costs close to zero (the cost of adding another person is low);
iii. high and increasing returns to the use of data (the more data you control, the better your product);
iv. and low distribution costs that allow for a global reach.
In addition to forced interoperability, the committee’s recommendations for countering these issues include creating special merger guidelines for digital platforms and giving special scrutiny to how factors such as product design and data control affect competitiveness.
How should policy makers react?
Naturally, the policy responses to these phenomena depend critically upon answers to some of the above questions. If concentration is growing but market power isn’t, the antitrust establishment may be doing exactly what it should. As Syverson points out, concentration itself is not an indicator of a problem.
If market power is expanding due to specific policy decisions, such as the threshold amendment to the Hart-Scott-Rodino Act, it may be necessary for policy makers to revisit, revise, or undo them. However, given that some policies relevant to competition, such as interest-rate levels, aren’t specifically antitrust issues, altering them in the service of healthier markets would have to be weighed against their effects on other areas of the economy.
But perhaps market power is growing, and not because of any single policy or set of policies. Perhaps it is growing because of a change in the nature of American commerce—such as the superstar-firm dynamic described by MIT’s Autor and his coresearchers. Monopolies have traditionally been studied as a microeconomic phenomenon, and policy responses to them tend to be individualized: the blocking of specific mergers or, in more extreme cases, the breaking up of specific companies. But if monopoly power is growing economy-wide, does that call for a broader and more fundamental response?
“The issue is, what’s the criterion that you use to declare something as noncompetitive in a way that’s easily comparable across industries?” Syverson says. Some have suggested that size—measured by total asset value, for instance—could be such a criterion, but “I think that is just taking a hatchet to a birthday cake,” he says. “It’s ridiculously blunt.”
If market power is so pervasive that it’s become a macroeconomic issue, Syverson says, that would require regulators to systematically work their way through markets one at a time, rather than make a policy adjustment that would affect all markets simultaneously.
Such an approach would be daunting for policy makers, but it is potentially critical. If markets are indeed becoming less competitive, that could mean not only paying higher prices for worse products, but also enduring lower wages, fewer innovations, and less business investment, as well as a smaller, slower-growing economy. It’s no surprise, then, that competitive markets are a matter near to the heart of most economists—and, perhaps increasingly, a salient part of politicians’ message to voters.