Jonathan B. Baker and Fiona Scott Morton challenge the interpretations of two new papers from Carl Shapiro & Ali Yurukoglu and Nathan Miller, which question economy-wide trends toward a rise in market power and, if any such trend has occurred, that it is due to lax antitrust enforcement.
Five years ago, we reviewed the economic evidence showing rising market power in the United States economy. Our conclusion was supported by multiple studies employing very different methods, each with different potential vulnerabilities. Notwithstanding the possibility that some of the evidence of rising markups had, in part, benign explanations other than increasing market power, we concluded that the overall picture was clear: “market power has been growing in the U.S. for decades.” This has been a general tendency; we did not claim that competition has lessened in every industry. We also pointed to evidence that antitrust was “falling short” and recommended ways to enhance competition by strengthening antitrust law, enforcement, and institutions.
Two new papers, one by Carl Shapiro and Ali Yurukoglu and one by Nathan Miller, question whether there has been an across-the-board rise in market power, based in part on research that appeared since we wrote in 2019. Both papers also suggest that economy-wide trends have not been the product of inadequate antitrust enforcement, though they each recommend strengthening antitrust policy on other grounds. For the reasons we discuss below, we continue to think market power is a substantial problem that has grown more important over the past half-century, in part because of insufficient antitrust enforcement.
Both papers adopt rhetorical frames that downplay the possibility of lessened competition by suggesting that growing market power cannot be important when industry trends also have other explanations. Shapiro and Yurukoglu “contrast” the “decline-in-competition” hypothesis with the “competition-in-action” hypothesis. Miller concludes that if market power has increased broadly across the economy, “technological change, rather than weak antitrust enforcement,” has been “the more important catalyst.” When writing analytically, however, both papers make clear that first, it is not possible to measure empirically which dynamic has been more important, and second, that these explanations are not in fact exclusive. This latter observation is the point we wish to highlight. The validity of the empirical findings referenced in these papers in no way rules out our view that insufficient competition enforcement has contributed importantly to an increase in market power. The two literature reviews are instructive but the findings they detail do not shed analytical light on whether competition policy has been too lax.
These papers argue that there is no empirical work proving that rising market power has been the product of lax enforcement and that the relevant data show patterns that could be explained by technological change. We agree. However, those patterns could also be due to insufficient competition enforcement. Lacking empirical methods to nail down the quantitative contribution of inadequate enforcement, competition economists can either decline to contribute to analyzing and addressing a pressing policy problem or use other tools to come to conclusions that can help guide policy. We prefer the latter approach.
Everyone agrees that market power has increased over time in many sectors of the economy. The authors of these literature reviews do not rule out a causal relationship between competition enforcement and economy-wide markups, either through their interpretation of empirical evidence or through economic theory. Instead, they emphasize the role of technological change in altering markups (price/cost margins). Given their rhetorical frames, that emphasis creates the impression that antitrust policy is not relevant or important. While the authors disclaim that conclusion, our fear is that those who gain from lax enforcement will use these writings to call into question the value of antitrust enforcement, thereby distorting the policy debate.
Put differently, competition policymakers cannot do anything about the consequences of technological change for market power beyond fostering competition in price, quality, and innovation. Accordingly, academics concerned about our market power problem who recognize the influence of technological change should understand the research findings these papers review as underscoring the importance of studying, discussing, and encouraging optimal antitrust enforcement, not as suggesting that such efforts are pointless.
We agree with a great deal of what our colleagues wrote. We share their methodological perspective: industries differ in many ways and industry-specific factors are important for understanding how firms compete. We agree with Shapiro and Yurukoglu that the available evidence on trends in concentration across markets is uninformative. We agree with both papers that methodological challenges make it hard to identify long-term trends in markups with precision. We nevertheless understand the recent literature to find that markups have in general increased economy-wide over several decades, and substantially so (albeit perhaps less than the original studies found). Given the invention and diffusion of the computer and digital technologies, which often raise fixed costs while reducing marginal costs, it would strain credulity if empiricists had not found that markups have risen. We also agree with the authors when they say that antitrust enforcement is important and needs to be strengthened regardless of how one looks at economy-wide trends.
Lax antitrust enforcement contributes to growing market power
However, we continue to believe antitrust enforcement contributes to insufficient competition in many industries for two reasons. First, economic studies have identified market power problems in many specific industries. A partial and unsystematic list focusing on product markets includes airlines (here and here), anesthesiology, brewing, cellular telephone service, dialysis, hospitals (here and here), online advertising, pharmaceuticals, rental housing, spectrum auctions, super–premium ice cream, soft drink bottling, and search engines. Market power in input markets (monopsony power) has been identified in markets involving college faculty, health care workers at hospitals, and workers in a range of occupations. In many of these examples, the harm is to consumers or workers who are less well off. The breadth of these industries and their importance—for example, the hospital industry is about 4% of GDP—contributes to our view that market power is a common problem in the U.S. economy. We recognize that all markets have not evolved the same way. In other examples, such as the cement and wholesaling industries, market power and efficiency have both increased or, as in the steel industry, market power has decreased. This heterogeneity is neither surprising nor a reason to question our view that market power has increased on average economy-wide.
Other evidence suggests that antitrust law and enforcement has not adequately deterred anticompetitive conduct. Many express cartels go undiscovered. Cartels may be the tip of an iceberg: tacit collusion is probably even more prevalent than explicit collusion because it is harder to prosecute and deter. Multiple studies demonstrate that firms have successfully hidden their anticompetitive mergers from antitrust authorities.
Second, antitrust rules were relaxed substantially beginning nearly five decades ago,. Since that time, the Supreme Court has, through its opinions, repeatedly signaled its preference for a non-interventionist approach to interpreting and applying the antitrust laws. That preference has been unmistakable, encouraging the generalist judges of the lower federal courts to interpret decisions as going well beyond their (potentially more defensible) narrow holdings.
The Court began in the late 1970s by opening the door to allowing procompetitive justifications for both vertical conduct and agreements between rivals concerning price. Although the Supreme Court has not decided a substantive merger case since the early 1970s, an influential opinion of the D.C. Circuit Court of Appeals did the same for horizontal mergers, including mergers in highly concentrated markets, through its receptivity to defenses based on showing that entry is easy or that sophisticated customers can protect competition by exercising buyer power. While that shift in approach was not inherently problematic, it facilitated a decades-long judicial movement led by the Supreme Court toward making antitrust law steadily more tolerant of market power and distrustful of antitrust enforcement. Those messages were reinforced and operationalized by the increasingly high burdens the Court placed on antitrust plaintiffs.
This dynamic began when the Court encouraged lower courts to evaluate the economic plausibility of antitrust claims and to throw out cases that make “no economic sense.” That mandate initially applied just to predatory pricing cases, but the Court soon suggested that tacit collusion in oligopolies is unlikely to be achieved or be stable. It later added that unilateral refusals to deal most likely benefit competition and that vertical conduct often poses no risk to competition. Cementing this welcoming attitude toward market power and even monopoly power, the Court asserted, remarkably and without support, that the opportunity to charge a monopoly price encourages innovation and growth, that any market power would be temporary, and that antitrust intervention against dominant firms often chills beneficial conduct.
To bolster its perspective, the Court made it harder for antitrust plaintiffs to succeed. When a judge questions a claim’s economic sense, the plaintiff must come forward with more evidence. When a defendant proffers a plausible procompetitive justification for an agreement or its own conduct, the Court has suggested, the restraint can be upheld regardless of whether it also harms competition, even when that harm is substantial. When the agreement is vertical, direct evidence of anticompetitive effects like higher prices is insufficient; the plaintiff must also prove market power. Even in a horizontal price-fixing case, the plaintiff must show that the agreement has “substantial” anticompetitive effect, and that defendants have market power, while the defendant need only proffer a procompetitive “rationale.” The Court justified this defendant-friendly approach to the rule of reason by fear of mistakenly condemning legitimate business arrangements.
These decisions bind and guide the lower courts. Hence their messages strongly influence the day-to-day functioning of antitrust enforcement, both public and private. Even though some lower court decisions like Microsofthave identified avenues for strengthening enforcement, a conservative Supreme Court is lurking, ready to pounce to prevent antitrust from stepping in to defend competition.
Against the background of the many troubling examples of industries where firms exercise market power, the many cases where the Supreme Court has signaled to lower courts that they should be reluctant to find antitrust liability, and the empirical evidence that antitrust does not adequately deter anticompetitive conduct, it is reasonable to conclude that lax antitrust enforcement is at least in part to blame for increases in market power in multiple industries.
Market power and technological change
Shapiro and Yurukoglu devote substantial attention to retrospective analyses of the effects of consummated mergers. This literature is hard to interpret: the mergers chosen for retrospective analyses are not randomly selected, the studies find varied effects, and none of the studies is a true experiment. However, it seems clear to us that many of the harmful mergers that have been identified in these studies could have been recognized at the time through detailed case-specific analysis, particularly in the hospital industry, but not just there. We think the retrospective analysis of close calls shows the same thing. Like other areas of antitrust enforcement, therefore, merger enforcement has been too lax.
We agree with our colleagues that technological progress often generates scale economies and more concentrated markets. As John Sutton has shown, this dynamic often occurs when firms adopt new technologies that involve higher fixed costs and lower marginal costs than before. During the past few decades, technological progress involving information technology has likely been the source of structural change in many sectors, not just in the information technology industries themselves. It is plausible that some firms within any industry took more advantage of the growing importance of scale economies than others, leading in some industries to the growth of “superstar” firms. The implication, which we think Shapiro and Yurukoglu would also draw, is that both margins and industry concentration would be expected to rise in many industries and on average overall.
Shapiro and Yurukoglu characterize this dynamic as competition-in-action, but it isn’t just that. Higher markups would not reflect greater market power if they simply represented temporary returns to the early adoption of information technology within otherwise competitive sectors. But they do reflect greater market power, at least in part, when the same investments that increase scale economies and concentration also deter entry and soften competition—as those investments often would be expected to do. As Miller also notes, scale-increasing technological change can result in fewer firms, thereby making coordination easier to sustain or reach. In consequence, tacit collusion could have become more common or more effective in many industries.
Miller emphasizes the importance of technological change for the long-term evolution of six industries that have been the subject of recent empirical research: consumer packaged goods, cement, wholesalers, steel, automobiles and airlines. Market power increased in four of these six industries but prices did not increase. In those four industries technological change led to reductions in marginal costs, but the firms did not pass through that cost reduction to their buyers. Miller views this observation, and the related observation that there has been almost no empirical relationship between markup changes and price changes in a wide range of industries over the past three decades, as indicating that technological change rather than weak antitrust enforcement explains any economy-wide increase in market power.
We disagree with Miller’s interpretation because competition would commonly be expected to lead firms to pass through at least some of their marginal cost reductions to their buyers. That doesn’t always have to happen, but it should happen often. Thus, the key policy question is not whether increased markups led buyers to pay more than they did before costs fell; it is whether buyers were harmed relative to a but-for world in which greater competition would flow from stronger antitrust enforcement.
The empirical studies that Miller reviews do not answer this question. Greater competition would be expected to encourage even greater beneficial technological change as well as to encourage firms to pass through those benefits to buyers. Of course, antitrust enforcement may not always be able to foster more competition when technological change leads industries to grow more concentrated and markups to rise. Regulation or other tools may be needed to achieve more competitive outcomes. In some industries, perhaps, the sunk investments associated with technological change and extent of the resulting differentiation among sellers would raise entrant minimum viable scale to the point where even perfect enforcement would not encourage entry. Understanding the potential and limits of antitrust enforcement in such settings is a pressing research question.
In addition, and more technically, our interpretation differs from Miller’s because the studies he references that attribute higher margins to lower costs rather than higher prices rule out by assumption the possibility that (tacit) coordination has become possible or more effective in those industries.
Decades of structural changes arising from technological change have combined with insufficient antitrust enforcement to increase market power in much of the economy. We continue to believe antitrust law and enforcement need to be strengthened to deal with it.
Authors’ Disclosures: Both authors consult for private plaintiffs and defendants as well as for governments. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
