The debates around antitrust reform have sometimes involved participants taking extreme positions. In some sense, this is good because it helps attract attention and there has been lots of attention from politicians, policy makers, academics, and the public. But it is not good if the debate winds up creating myths that are unhelpful in forging a path forward. I am convinced that there is a sufficient body of evidence to establish that, although there are many improvements that can be made in antitrust doctrine and enforcement, the claims by a growing number of academics, politicians, and government officials (often referred to as “Neo-Brandeisians”) that antitrust needs to be radically redirected and that the core principles that have guided it for the past half century should be jettisoned are wrongheaded and would lead to undesirable policy outcomes. My goal in this short essay is to explain a few key points that I hope many will see as obvious. I first start out with some basic theoretical/philosophical observations, and then move on to empirical ones. After setting a common theoretical and empirical background, I discuss whether there is a need for change and, if so, what change, for the major antitrust doctrines concerning cartels, mergers and exclusionary behavior. I then go on to discuss some possible improvements in how economics can be used in antitrust matters, as well as other ways in which we can better obtain the benefits of competition without, in effect, throwing the baby out with the bath water.
By Dennis W. Carlton[1]
I. INTRODUCTION
The debates around antitrust reform have sometimes involved participants taking extreme positions. In some sense, this is good because it helps attract attention and there has been lots of attention from politicians, policy makers, academics, and the public. But it is not good if the debate winds up creating myths that are unhelpful in forging a path forward. I am convinced that there is a sufficient body of evidence to establish that, although there are many improvements that can be made in antitrust doctrine and enforcement, the claim by a growing number of academics, politicians, and government officials (often referred to as “Neo-Brandeisians”) that antitrust needs to be radically redirected and that the core principles that have guided it for the past half century should be jettisoned are wrongheaded and would lead to undesirable policy outcomes.
My goal in this short essay is to explain a few key points that I hope many will see as obvious.[2] I first start out with some basic theoretical/philosophical observations, and then move on to empirical ones. After setting a common theoretical and empirical background, I discuss whether there is a need for change and, if so, what change, for the major antitrust doctrines concerning cartels, mergers and exclusionary behavior. I then go on to discuss some possible improvements in how economics can be used in antitrust matters, as well as other ways in which we can better obtain the benefits of competition without, in effect, throwing the baby out with the bath water.
II. BASIC THEORETICAL/PHILOSOPHICAL PRINCIPLES
Let me explain some basic principles in this section- principles that seem compelling to me, but that seem likely to generate vigorous disagreement from today’s Neo-Brandeisians.
Principle 1: Vilification does not advance the debate about the usefulness of economics, and economic principles should form the foundation for sound antitrust policy.
Although I understand the debating value of discrediting one’s opponent, I think little is gained by antitrust’s critics from demonizing Robert Bork, the Chicago School, or economists (and economics) in general. What people mean when they talk about “Chicago” is often unclear, or worse, a serious mischaracterization. I define “Chicago School,” or at least the Chicago School related to economics, as the rigorous application of microeconomic theory combined with empirical evidence to test the theory. I am sure some will disagree with that definition, but that is what my experience over the last 45 years on the faculties of the business school, economics department, and law school at The University of Chicago has taught me.
Kovacic (2020) has made clear the contribution of scholars from many universities, including Harvard, to the use of economics in order to understand and help guide antitrust policy.[3] The principle that economics has something useful to say about antitrust policy should not be controversial. Economists, especially industrial organization economists, have for many years been studying how firms do or do not compete. No doubt there is much more to be learned, but it is also true that there has been much that already has been learned. Ignoring what economists have learned would be throwing away that knowledge and not likely to produce desirable policy. Indeed, the poorly decided antitrust cases of 40 or more years ago well illustrates the danger of ignoring economics.
To figure out how antitrust policy should be formulated, one needs to have an objective in mind for antitrust policy. Although I understand that others may want to pursue goals such as the preservation of small firms, there is only one goal that makes sense to me for a policy that is presumably focused on competition.
Principle 2: The objective of antitrust should be to enable firms to compete with each other.
But, of course, one must define what “ to enable firms to compete” means. I would define it as a process that produces certain desirable economic outcomes. When firms compete with each other, certain desirable outcomes generally emerge. When firms compete with each other, firms have an incentive to minimize their costs, and competition forces them to pass on those low costs to consumers. Inefficient (i.e. high-cost) firms will find it hard to survive since the price that they can charge cannot cover their costs. This elimination of inefficient firms is not a “harm to competition” that should be prevented since the resulting outcome of the competitive process produces benefits to both firms and consumers. Innovative firms have an incentive to develop a new product to benefit consumers because the firms can reap profits from the sales of that product. Even though those firms may earn high profits, consumers are better off. Those high profits provide an incentive for firms to enter and drive prices of the new product down which further benefits consumers. Firms competing for workers raise wages. These desirable economic outcomes—low prices. innovative products, and high wages are not the result of some magical theorem in economics, as the next principle explains.
Principle 3: Economic theory does not prove that competition necessarily produces the socially optimal prices or mix of goods. For example, in the presence of uncertainty, market power, imperfect information, a lack of complete markets, externalities or high transaction costs- the outcome of the competitive process can certainly depart from the welfare maximizing ideal of introductory economics textbooks
Though Principle 3 is not (or should not be) controversial even among the most fervent advocates of laissez faire, it is a grievous error to jump from it to the conclusion that government intervention, in this case through antitrust policy, can be expected to improve outcomes. Many years ago, Harold Demsetz made this point quite clearly when he explained what has become known as the “Nirvana fallacy.”[4] Just because in an ideal world-which does not exist- one can theorize that matters could be improved by an omniscient planner, that does not mean that government intervention will actually achieve these desired outcomes. Indeed, the empirical evidence indicates that government regulation or government intervention in markets to “improve” outcomes often makes matters worse.[5] Some have labelled this phenomenon as “government failure,” which creates harms to the economy just like the more often cited-to “market failure.”
The preference of economists for not interfering in markets comes from the empirical recognition that desirable outcomes typically arise when firms compete in the real world even if that world is not the one of introductory economics textbooks. This does not mean that regulation or government intervention is never needed, but that one needs to ask not only whether some form of government intervention might be able to improve outcomes in theory, but also whether there is good reason to believe that it will do so in practice.[6] So this leads to Principle 4.
Principle 4: Absent special circumstances, competition among firms is desirable.
Recently there has been a lot of discussion about whether to broaden the goals of antitrust policy beyond merely the desirable economic effects of competition on prices, wages, and innovation that I have just discussed. This leads to Principle 5.
Principle 5: Using antitrust policy to pursue goals other than achieving the desirable economic effects of competition—however worthy those other goals may appear to be– would be a mistake.
There are many effects that arise as a consequence of competition even when competition is serving to bring about desirable outcomes such as low prices, high wages, and innovative products. For example, competition can lower the wealth of a firm’s owners because it leads to lower prices. Should it matter for antitrust policy whether the owners are “deserving” or “greedy”? If competition drives down prices of say, cigarettes, should the health hazards from additional smoking be considered in formulating antitrust policy? If so, should we let the cigarette firms form a cartel to reduce smoking through high prices? (Wouldn’t it be better not to subvert competition by allowing a cartel and instead have competition but tax the firms?) Suppose competition encourages firms to use cheap fuel that pollutes, or to produce and sell consumers more and more goods and services, potentially contributing to global warming? Should that matter for antitrust policy? Should it matter whether some inefficient firm employs lots of workers and needs protection from competition to survive?
It is beyond debate that each of these effects can raise legitimate policy concerns, but those concerns, in my view, are not well- suited to antitrust policy. Can any of these effects be deemed a “restraint of trade”? Antitrust policy is premised on the central idea that competition is good, not bad. It is the job of other government policies to deal with all these other policy issues that competition might or might not affect. If antitrust policy is turned into a tool to achieve every desirable goal imaginable, then there is no guidance at all as to what the tradeoff among these differing goals should be.
A policy that considers everything in its decision making is a policy doomed to fail and one that is likely to turn antitrust policy into a political weapon to be wielded against disfavored firms or individuals, untethered from any well- defined principle, or to lead to court decisions that depend on the idiosyncratic preferences or political leanings of the judges and juries. I suspect that the current desire to use antitrust policy to pursue goals other than the protection of competition comes from the political gridlock currently facing the United States in which Congress seems unable to accomplish much with respect to these other goals. Although it is understandable why some find this gridlock frustrating, diverting antitrust from its central focus of competition would likely make matters worse.[7]
There can still be complications in understanding the effects of certain conduct on consumers (or input suppliers such as workers) even if one focuses on just the economic effects that I have discussed. Consider a merger and for simplicity, focus on price. There can be winners and losers among consumers after the merger if the merger affects prices charged to different consumers differently. Suppose, for example, that an airline merger leads an airline to move capacity from one route to another, leading to much lower fares and greatly expanded output on one route but slightly higher fares and slightly reduced travel on the other. How should that be handled?
One could try to weight the harm to one group against the gain to another with the weights depending on how “deserving” each group is deemed by the policy maker. A less discretionary, and far more practical, approach would be to calculate the net gain overall to all consumers (or, more generally, to all affected parties including input suppliers such as labor) and determine whether it is positive.[8] This is the approach for example that the DOJ typically follows in evaluating airline mergers. That seems sensible to me and avoids placing arbitrary “value weights” on different consumers.[9]
Let me end my discussion of basics with a topic that has received disproportionate attention. That topic is the difference between total and consumer surplus.
Principle 6: The difference between total surplus and consumer surplus is of little practical import for U. S. antitrust policy.
There is of course a difference between the two concepts and economists trained in cost benefit analysis are likely to naturally use total surplus. [10]The difference between the two depends on the profits earned by firms. Consumer surplus is the roughly triangular area bounded above by the demand curve, below by the horizontal line at price, and by the vertical price axis. Producer surplus is the roughly triangular area bounded above by the horizontal price line, below by the supply (marginal cost) curve, and by the vertical price axis. Total surplus is the sum of consumer plus producer surplus. An interference with competition can result in a deadweight loss, a loss of total surplus.
Although the famous Williamson 1968 paper makes clear the tradeoff between efficiency and deadweight loss (using total surplus), there is no reason one could not also use consumer surplus in that analysis.[11] An efficiency in that setting would exonerate conduct only if consumers were better off. And that is how U.S. antitrust authorities have typically dealt with conduct such as a merger that could harm consumers.
Some claim that the use of total surplus to evaluate antitrust conduct is a huge mistake because it credits the effects on firms’ profits, but the truly huge mistake is thinking that this claim is of any practical import. The reason is that I have rarely (never?) come across a U.S. case where the use of total versus consumer surplus matters to a policy outcome! Specifically, I am unaware of any U.S. case where the antitrust authority allowed some activity such as a merger or exclusionary behavior because it generated huge efficiencies and simultaneously harmed consumers.[12] If I am correct, the difference between consumer and total surplus is a good topic for students to think about to test their knowledge, but it is a topic that has little practical import.
For those who adopt a consumer surplus standard because they are uncomfortable adopting a total surplus standard that could allow a firm to profit even if consumers are harmed, but are still troubled by prohibiting a transaction that increases total surplus, there may be a way to eliminate this tension. As long as transaction costs are low enough, even if the policy standard is consumer surplus, then a merger, for example, that generates efficiencies and increases total surplus but harms consumers, can be transformed through non- linear pricing into one that increases consumer surplus. Hence, it is possible for that merger to be approved, assuming the non-linear pricing can be contractually guaranteed.
Those who claim that antitrust should pay attention only to consumer surplus are clearly mistaken, and would be advocating policies in tension with what are likely to be their preferences for greater competition by purchasers of labor services. Economists decry harm to competition on either the buying side or selling side. For example, a cartel of consumers that lowers price benefits consumers but causes a harm of no less economic significance than the harm caused by a cartel of sellers. One is monopsony, the other is monopoly. Both harm the competitive process, restrict output and create deadweight loss. [13]
Now that I have gone through what I hope most will regard as compelling basic principles, let me state a few important empirical facts about which I hope there is general agreement.
III. IMPORTANT EMPIRICAL RESULTS OF RELEVANCE TO ANTITRUST
Increased industry concentration is often good. Demsetz, long ago, pointed out that the reason that some industries see increased concentration is because the most efficient firms grow and replace inefficient ones.[14] Recent empirical findings bear this out. Autor et al. (2020) and De Loecker et al. (2020) both find that although the median firm in an industry has seen little change in its markup, the biggest firms in each industry have seen large increases in their markups.[15] Research shows that these large firms are typically the most efficient in the industry so that productivity growth and increased industry concentration go hand in hand. Surprisingly, the price in these industries has not fallen, consistent with higher margins for the most efficient firms.[16] De Loecker et al. (2020) and others find that the average price cost margin has increased since 1980.
Prices sometimes rise and sometimes fall post- merger. There is significant evidence that in many (though by no means not all) consummated mergers that presented close calls about whether an agency should sue, prices have risen nontrivially post-merger. Importantly, however, it can be hard to identify those adverse cases ex ante. It is also true that in many cases, prices do not rise and instead fall post-merger. The fact of many examples of price increases, together with evidence that in many (though by no means all) such mergers, remedies did not work as effectively as hoped, have induced calls for more strenuous antitrust treatment of mergers.
However, no one, as far as I know, has figured out the extent to which more stringent antitrust policy toward merger activity might inhibit actual efficiency enhancing mergers from even being attempted. (Nor, for that matter, has one been able to determine whether mergers that were successfully challenged, or were abandoned due to Agency opposition, sacrificed efficiencies and benefits that might have redounded to consumers). Without knowing the answer to that question, it is hard to figure out an optimal enforcement policy. Perhaps the best one can say is that if one did not expect to see such evidence of many post-merger instances of price increases, then one should increase the bar for allowing mergers—particularly, perhaps, in industries where this has formed a pattern–but it is an area deserving of more research.
The claim that industries in the U.S. have become much more concentrated is overstated. First, to credit the evidence put forth by those claiming that markets have become much more concentrated over time, one must accept that the national data using NCAIS or SIC classification define an antitrust market, even though the data ignore (among other things) imports, local markets and demand substitution from products made with different technologies. Second, if one looks at, say manufacturing, one would see that although there has been an increase in concentration, it is not true that that level of concentration is very high (say HHI>1500) for the vast majority of industries. See my 2019 CPI article summarizing this evidence.[17] Finally, growing industry concentration need not be the result of any failure of antitrust policy towards mergers (or more generally towards anticompetitive conduct). Organic growth of efficient firms surely matters in the explanation for increasing concentration.
Entry is not easy in many industries. In a classic paper on entry, Dunne, Roberts & Samuelson (1988) showed how difficult it was for a firm to enter into a new manufacturing industry, defined as 4 digit SIC code.[18] The cumulative failure rate for an entrant was very high on average after 5 years (around 60 percent) and 10 years (around 80 percent). Subsequent studies have shown how hard it is on average for entrants to build up demand over time. Any reliance on a general presumption that entry will quickly remedy lax antitrust policy or enforcement seems misplaced.
IV. WHAT AREAS OF ANTITRUST, IF ANY, NEED CHANGE?
With the common theoretical and empirical background of the prior sections, I now evaluate whether any areas of antitrust need change. My view is that although antitrust reasoning by the courts can and should be improved in several areas, there is no need for new antitrust legislation. The common law system seems adequate to deal with the deficiencies in the major antitrust areas I discuss. But I am not a lawyer, so I confine my discussion mainly to economic matters, focusing on where existing antitrust doctrine could be improved.
Cartels on the buying side or selling side are bad.Naked agreements among independent firms to restrict output and raise price and naked agreements among independent firms to not compete for inputs such as labor are bad. It is hard to conceive that this statement would be controversial. The DOJ has used its leniency program combined with its ability to criminally charge conspirators with the threat for fines and jail sentences to create incentives for firms to reveal the existence of cartels. More recently, it has pursued firms engaged in no- poach agreements to restrict their competition for workers.[19] Both seem like correct policies to pursue. It is of course true that there are always subtleties that generate exceptions. For example, some agreements to restrict labor mobility may be justified by a need to protect trade secrets or to provide firms with incentives to train workers, while some agreements to fix price (e.g. BMI) may be justified in order to create a new product.[20] But courts have shown that they can deal with such exceptions successfully. Exceptions should not be used to defang a per se prohibition on naked cartels on either the buying or selling side.
Horizontal mergers that create market power and raise prices or lower wages or reduce innovation are bad. Again, it is hard to believe this is controversial. Where the controversy arises is not on the theoretical side but on the empirical side. It is hard enough to predict when a merger might increase price but can be extraordinarily hard to predict which firms will produce and market successfully a new product in the future. But predicting which firm will have the next new great product absent a merger is the task that is required if one wishes to stop mergers involving nascent competitors. One might simply wind up preventing small firm acquisitions, leading to a decline in small firm start- ups.
I am skeptical in general of the claim that, other than in specific cases such as pharma—where one often knows the drugs that are actually under development by potential entrants—one can make very accurate predictions of who will invent what in the future, especially in rapidly changing industries. That suggests that aggressive actions to stop mergers in a general rather than targeted attempt to prevent acquisitions of “nascent “competitors is likely to stop acquisitions of start-ups and thereby deter the creation of start- ups that hope to be acquired, with little or no positive benefit to competition. Unless evidence clearly establishes both that the “nascent” competitor is a likely entrant AND that it is somehow unique (i.e. that there are not a number of others who can serve the same role), a broad policy of attacking such acquisitions seems likely also to impede the ability of firms to obtain desirable synergies through merger, while at the same time reducing the incentive for start- ups to form in the first place.
Some commentators have questioned whether horizontal mergers generate any (merger-specific) efficiencies. If there are no efficiencies then there is no harm in stopping horizontal mergers. As one might expect, there is evidence of efficiencies from specific mergers and also evidence showing that mergers do not always create efficiencies. I agree that it is important to figure out the efficiencies from any horizontal merger and this is a current area of research. But anyone who has ever looked at the number of mergers over time in what looks like a relatively competitive industry would wonder why there are so many transactions if not driven by some efficiency motive. So adopting a position that horizontal mergers generate no efficiencies strikes me as extreme. There is a debate as to where the legal burden should be on showing harm versus showing an efficiency. My economic view is that an economist should give the best prediction as to the net harm or net benefit that a merger will create. It is not obvious to me why either gross harm or gross efficiency should receive more attention than the other.
As already discussed, the empirical evidence does show that many consummated mergers have led to price increases. The evidence also shows that many mergers did the opposite and led to lower prices. It is the ability to predict price increases ex ante that should be the focus of research. But, assuming that our analytic methods have some validity, and in light of the evidence on adverse price impacts of many mergers, there is no reason to be timid in attacking mergers whose predicted price effects are positive, absent clear evidence of offsetting benefits of some sort.
There are at least three research areas worth mentioning in regard to horizontal mergers. First, a literature on common ownership has shown that among publicly traded firms, it is often the case that a few large institutional shareholders (e.g. Vanguard index fund of the total stock market) own shares in all the major rivals in some industry. Some claim that this common ownership creates an incentive for the common owner to influence the rivals in an industry to compete less vigorously against one another in order to raise industry profits. There are many flaws in this claim as a matter of theory – for example, an increase in the price of one product will raise the costs to firms in other industries that use this product and that would depress the value of stocks held by the institutional shareholders in a total stock market fund- but there may be some theoretical plausibility to the theory under some circumstances.[21] Although the empirical demonstration of the mechanism of harm is still a subject of debate, as well as whether common ownership has any effect at all on prices, this research bears watching. Common ownership is similar enough to overlapping directorates to make it worthwhile to keep this topic on the radar screen when evaluating horizontal mergers.
Second, there is a growing literature on the existence of market power in the hiring of labor. Ashenfelter et al. (2021) discusses how labor economists have been amassing evidence that labor markets are less competitive than once thought.[22] That means that mergers that reduce the number of firms that hire a specific type of labor could harm competition for those workers. For example, in a merger of hospitals, there could be harm in the labor market for nurses who suffer lower wages and employment as a result of the merger. Such harms from mergers should not be ignored.
A third area of research deserving more attention is coordinated behavior. Much research in merger analysis over the past couple of decades has focused on unilateral behavior, in which the merging firm takes account purely of how it can coordinate the pricing of its products with those of the firm it is merging with. But much less attention has been focused on how a reduction in the number of firms in an industry could improve oligopolistic coordination, in which a few firms do not aggressively compete against each other because they realize such competition will ultimately make them all worse off, once rivals react. This area of research deserves more attention
Vertical mergers raise different issues than horizontal mergers and raise competitive concerns only under special circumstances. Unlike a horizontal merger which combines substitute products, a vertical merger combines complements. The combination of complements, assuming market power in their provision, creates an efficiency by eliminating double marginalization (or more generally the Cournot complements problem). It can also create a harm by creating an incentive for the merged firm to raise the price of an input to a rival. It is the combined effect of these forces that one must examine in order to figure out whether a particular vertical merger is desirable. The empirical literature on vertical mergers is more limited than that on horizontal, but the results generally seem to show no large harmful effects. Indeed, a detailed study of the only litigated vertical merger case in over 40 years shows that there were no harmful effects from a vertical merger (ATT/TW) that the Department of Justice tried unsuccessfully to block. (See Carlton et al. (forthcoming).[23])
I see no reason, at least from an economic viewpoint, why current antitrust law and practice are not adequate to handle both horizontal and vertical mergers. As evidence mounts from retrospective analyses of past mergers, presumptions can change, but the economic methods currently in use seem sufficiently flexible to handle new situations that might arise. If presumptions do change so that antitrust authorities investigate and/or challenge more merger cases, then it follows that the agencies must have expanded budgets to carry out their duties I discuss below some suggestions to help make the evolution of presumptions more accurate.
Section 2/ Exclusionary conduct cases are harder to discuss because the possible conduct is so varied. But, as I have explained elsewhere,[24] there are at least two areas of antitrust law that could be substantially improved from an economic perspective. One area is the tie-in sale doctrine. A tie-in sale is often no more than the efficient bundling of product characteristics together for efficiency. A car is a tie-in of a motor, wheel and tires etc. Creating liability for engaging in such a tie would be obviously foolish. But tie-ins can serve other purposes. One standard reason for tying is to meter usage so that heavy users pay more than light users. This is a form of price discrimination and generally has no effect on the competitive process. Indeed, if a monopolist engages in the behavior, there can be, by assumption, no competition among rivals to harm.
In contrast, if a firm has market power in product A, then a tie of product A with B can, under some circumstances, lead the firm to acquire market power over consumers of product B. The key to understanding whether a tie results in competitive harm is identifying whether the tie enables the firm to gain market power over a group of consumers over whom it would not otherwise have market power. As far as I can tell, current antitrust doctrine on tie-ins fails to pay sufficient attention to the different reasons that firms engage in tying and fails to identify clearly those ties that cause competitive harm and those that do not.
The second area where improvement is needed involves so called “two-sided markets”. The Court’s decision in Amex ruled that the antitrust standards for evaluating the legality of exclusionary conduct depend on whether the market under analysis is two-sided or one-sided.[25] Two-sided markets arise when there are indirect network externalities so that the total number of consumers on one side matter to the other side and perhaps vice-versa. Examples include a dating bar , a transaction platform, and a newspaper that serves readers and advertisers. My article with Ralph Winter explains why the Court’s decision in Amex was a mistake.[26] The Amex decision has led to subsequent cases in which defendants are exonerated even though the findings indicate competitive harm. The mistake in Amex was an overreliance on the formalism of market definition. I will return to this topic below when I discuss market definition and the ignoring of evidence of actual harm.
Finally, one reason for much of the conduct that is often attacked under Section 2/exclusionary conduct is to create an incentive to provide sales effort when there is a free rider problem where a firm other than the one providing the sales effort can enjoy the benefit of the sales effort. Consider, for example, a manufacturer whose product is distributed by retail firms. That manufacturer may want to use exclusive territories in order to enhance sales effort by its distributors. In the absence of exclusive territories, one retail firm providing selling effort such as advertising may increase the sales of its rival across the street and reap no benefit from its sales effort.
With exclusive territories, the sales effort is likely to generate increased sales in the territory of the firm providing the sales effort, thereby rewarding the firm providing the sales effort and eliminating the free rider problem. But digital technology has greatly expanded the ability to monitor who is providing sales effort and to whom and whether that sales effort ever ultimately results in a sale. If that sales effort can be monitored, that lessens the need to have vertical restrictions since the manufacturer can use these other methods of monitoring to directly reward the provider of the sales effort that results in a sale. In other words, the ability to write and monitor more complete contracts lessens the need for vertical restrictions.
Over- reliance on market definition. It is wrong to think that market definition is anything other than a crude construct that may help one think about a matter, but it is no alternative to more detailed analysis. It is a place to begin, not end, an analysis. Unfortunately, because of the heavy reliance on market definition in antitrust cases, my experience is that defendants want to argue that the market is broad so as to be able to say that shares are so low that there can be no harm while plaintiffs want to do the reverse. Although reasonable as an initial simplification, any notion that products are either “in” or “out” of a market, rather than that they exert differing levels of competitive influence on the sale of another product is a simplification that can lead to error if relied on to the exclusion of direct evidence. For example, suppose one can show that certain conduct had the sole effect of raising price or making entry more costly. Then, arguing that the market is broad and demanding that a case be dismissed because of that, is using market definition to ignore the relevant economic evidence. That is exactly what happened in American Express, discussed earlier. I would caution against over-reliance on market definition to decide cases and use it instead as a crude measure – at most a first step to determine whether a further analysis is plausibly needed. But if the evidence shows that the effect of some conduct is to harm competition, using market definition instead of using the direct evidence makes little sense.
Process improvements in antitrust cases: I have been involved in many antitrust matters in the U.S. and in other countries. At trials in the U.S., especially jury trials, I find that economic evidence is not well presented in the sense that the experts often testify far apart in time from each other and use complicated models, so it is often hard for a judge or jury to figure out what the economists are saying and to identify why the experts come to different conclusions. I have found that a “hot tub” in which opposing experts appear at the same time can improve the presentation of evidence. I would recommend that hot tubs should be more widely employed, with the goal of helping explain to judges ,arbitrators or a jury why the two opposing experts disagree. Were they told to make different assumptions? Do they use the same assumptions but different models? Are they simply repeating what some company witness told them? Judges, arbitrators or jurors should be able to ask questions of each expert.
In matters before the federal agencies, an open dialogue between the opposing economists can help clarify points of disagreement. I have noticed a reluctance sometimes to let the economists talk to each other for fear of losing some litigation advantage. That is not a good way to run a transparent government, in my view. If the attorneys are worried about losing some litigation advantage, they could enter a stipulation that the economists can talk to each other openly but nothing communicated can be used in a trial should litigation ensue.
Finally, more retrospective studies of merger and other antitrust decisions are needed. This involves not only asking whether price (or wages or innovation) rose or fell after a merger (or some other conduct) was allowed or not allowed, but also asking what the models of the opposing experts predicted at the time the merger (or other conduct) was being evaluated. Who was right and why or why not? The failure to examine what the opposing economists at the time were saying impedes any attempt to improve the economic analysis of mergers or other cases. I would be in favor of a rule that said that the expert reports in either an adjudicated case or in an investigation that does not go to trial should be made public subject to redaction to preserve confidentiality. That means the models’ predictions (though perhaps not all the underlying data) can be evaluated. Only then will one be able to figure out which type of models best predict outcomes.
Promoting competition: Both the FTC and DOJ have an important role to play not just in using the antitrust laws to preserve competition, but in explaining to state legislatures and other regulators and political bodies why competition is desirable and should not be thwarted. One area where there could be substantial benefits to enhancing competition is from the elimination of the many barriers to competition that states have erected to protect various industry groups. For example, many states restrict the sale of automobiles except through franchises, many states inefficiently restrict how franchisors can deal with franchisees, many states restrict competition in the distribution of alcoholic beverage or soft drinks, state licensing boards often restrict who can compete in provision of medical services, certificate of need restrictions in a state impede entry into hospital markets, and there are an incredibly large number of state licensing restrictions on industries ranging from flower arranging to haircut provision, to mention just a few areas.
Even if sometimes such regulations can promote economically desirable outcomes, the danger of restricting competition should be taken into account. Unfortunately, it does appear that state legislatures are often unable to resist powerful interest groups who wish to protect themselves from competition. Any action on the part of the government agencies, including information campaigns that would hopefully constrain the ability of states to create barriers to competition, would be desirable and consistent with the central theme of antirust policy that competition is desirable.
[1] David McDaniel Keller Professor of Economics, Emeritus, Booth School of Business, University of Chicago, and National Bureau of Economic Research. I frequently consult on antitrust matters through Compass Lexecon. I thank Ken Heyer, Mark Israel, Daniel O’Brien, Gregory Pelnar, Sam Peltzman & Theresa Sullivan for helpful comments.
[2] I have explored many of these topics in depth in prior academic articles and my textbook. I refer the readers to some of that work for more detailed discussions. See, e.g. Dennis W. Carlton & Jeffrey M. Perloff, Modern Industrial Organization (4th ed. 2005); Dennis W. Carlton, Does Antitrust Need to be Modernized, 21 J. ECON. PERSP. 155-176 (2007); Dennis W, Carlton, Market Definition: Use and Abuse, 3 COMP. POL. INT. 3-27 (2007); Dennis W. Carlton, Transaction Costs and Competition Policy, 73 INT. J. IND. ORG. 102539 (2020); Dennis W. Carlton & Ken Heyer, The Revolution in Antitrust: An Assessment, 65 ANT. BUL. 608-627 (2020); Dennis W. Carlton & Mark Israel, Effects of the 2010 Horizontal Merger Guidelines on Merger Review: Based on Ten Years of Practical Experience, 58 R. IND. ORG. 213-234 (2020).
[3] William E. Kovacic, The Chicago Obsession in the Interpretation of US Antitrust History, 87 U. CHI. L. REV. 459-494 (2020).
[4] Harold Demsetz, Information and Efficiency: Another Viewpoint, 12 J. LAW & ECON. 1-22 (1969).
[5] See, e.g. NANCY ROSE (ED.), ECONOMIC REGULATION AND ITS REFORM (2014), and Clifford Winston, Back to the Good—or Were They the Bad—Old Days of Antitrust? A Review Essay of Jonathan B. Baker’s The Antitrust Paradigm: Restoring a Competitive Economy, 59 J. ECON. LIT. 265-284 (2021).
[6] I note that even when antirust intervention occurs, it rarely takes the form of explicit regulation such as price setting.
[7] Though as a theoretical matter, taking “everything” into account and calculating social welfare sounds reasonable as a way to make decisions, my view is that it is not practical when agencies, judges and juries have limited expertise. One could try to infer the trade-offs among various goals by empirically examining government decisions, but I see that as introducing more uncertainty into the decision process and I doubt it would be appealing to critics who already complain that even the economic models in use today, which do not take “everything” into account, are too complicated. A wholesale restructuring of regulation, antitrust and social welfare policies into one “super” agency that takes everything into account may be appealing in theory but, from my viewpoint, not practical, at least not now. In any event, it is way beyond what critics of antitrust are asking for and beyond the scope of this paper. I note that the optimal structuring of government agencies is an important topic for research.
[8] Rather than calculating economic gains and losses to each consumer and aggregating across consumers, one could adopt the simpler (and less accurate) procedure of asking what happens to total output and average price.
[9] I note that anticompetitive conduct in one market can also cause harm in related markets (e.g. for complementary goods) where market power exists and that deadweight loss should matter to antitrust policy decisions.
[10] There has been utter confusion of language because Bork used the term “consumer welfare” when he likely meant total surplus.
[11] Oliver E. Williamson, Economies as an Antitrust Defense: The Welfare Tradeoffs, 58 AM. ECON. REV. 18–36 (1968).
[12] But see Ralph Winter & Roger Ware, Merger Efficiencies in Canada: Lessons for the Integration of Economics into Antitrust Law, 61 ANT. BUL. 365-375 (2016).
[13] Monopsony in, say, the labor market leads to a lower wage and less employment than under competition. This reduction in employment leads to reduced output and higher output prices.
[14] Harold Demsetz, Industry Structure, Market Rivalry and Public Policy, 16 J. LAW & ECON. 1-9 (1973).
[15] David Autor, David Dorn, Lawrence Katz & Christina Patterson, The Fall of the Labor Share and the Rise of Superstar Firms, 135 Q. J. ECON. 645-709 (2020); and Jan De Loecker, Jan Eeckhout & Gabriel Unger, The Rise of Market Power and The Macroeconomic Implications, 135 Q. J. ECON. 561-644 (2020). For additional discussion and citations, see Dennis W. Carlton, Some Observations on Claims that Rising Market Power Is Responsible for U.S. Economy Ills and that Lax Antitrust is the Villain, 2(1) CPI ANTITRUST CHRONICLE 10-21 (2019).
[16] See, e.g. Sam Peltzman, Productivity, Prices, and Concentration in Manufacturing: A Demsetzian Perspective, 65 J. LAW & ECON. S121-S153 (2022).
[17] See, Carlton, supra note 14. See also C. Lanier Benkard, Ali Yurukoglu & Anthony Lee Zhang, Concentration in Product Markets (NBER, Working Paper No. w28745, 2021).
[18] Timothy Dunne, Mark J. Roberts & Larry Samuelson, Patterns of Firm Entry and Exit in U.S. Manufacturing, 19 RAND J. ECON. 495-515 (1988).
[19] I would be puzzled by any claim that either economists don’t care about monopsony or that the antitrust laws do not apply to buyer cartels. I attribute the lack of more cases regarding monopsony to the only relatively recent empirical findings showing that monopsony power, especially in labor, is more prevalent than previously thought. I discuss these more recent findings later.
[20] Broadcast Music, Inc. v. CBS, 441 U.S. 1 (1979).
[21] For a critique of the literature on common ownership, see, e.g. Daniel P. O’Brien & Keith Waehrer, The Competitive Effects of Common Ownership: We Know Less than We Think, 81 ANT. L.J. 729-776 (2017).
[22] Orley Ashenfelter, David Card, Henry Farber & Michael Ransom, Monopsony in the Labor Market: New Empirical Results and New Public Policies, (NBER, Working Paper No. w29522, 2021).
[23] I was the expert for ATT. Dennis Carlton, Giorgi Giozov, Mark Israel & Allan Shampine, A Retrospective Analysis of the AT&T/Time Warner Merger, J. LAW & ECON. (forthcoming). See also Carl Shapiro, Vertical Mergers and Input Foreclosure: Lessons from the AT&T/Time Warner Case, 59 R. IND. ORG. 303-341 (2021).
[24] See Dennis W. Carlton & Ralph A. Winter, Vertical Most-Favored-Nation Restraints and Credit Card No-Surcharge Rules, 61 J. LAW & ECON. 215-251 (2018); Dennis W. Carlton & Michael Waldman The Strategic Use of Tying to Preserve and Create Market Power in Evolving Industries, 33 RAND J. ECON. 194-220 (2002); Dennis W. Carlton & Ken Heyer, Extraction vs. Extension: The Basis for Formulating Antitrust Policy Towards Single Firm Conduct, 4 COMP. POL. INT. 285-305 (2008). The classic paper is Michael D. Whinston, Tying, Foreclosure, and Exclusion, 80 AM. ECON. REV. 837-859 (1990).
[25] Ohio v. Am. Express Co., 138 S. Ct. 2274 (2018).
[26] Carlton & Winter, supra note 22.
