larry white, is a professor of economics at New York University’s Stern School of Business.
Illustrated by kotryna zukaustaite
Published January 24, 2021
This is antitrust’s moment, with Big Tech in the crosshairs. Apple is already in court defending its app-store pricing and practices against challenges by game makers and disgruntled consumers. The Federal Trade Commission (one of Washington’s two antitrust enforcement agencies) and a covey of state attorneys general have charged Facebook with a slew of violations. And, arguably most important, the Justice Department is suing Google, the holy of holies for tech worshipers. That suit aims to change practices that are allegedly designed to sustain Google’s dominance in online search and advertising.
Why all this attention now? For one thing, the so-called digital platform companies have grown so large and so prominent that they seem omnipresent. Amazon (with some 800,000 employees) is the second-largest employer in the United States (behind only Walmart, at 2.2 million worldwide, which is racing to become an online megastore, too).
As of last fall, the five largest U.S. companies by market capitalization — Apple, Amazon, Microsoft, Facebook and Google’s parent, Alphabet, with a combined value exceeding $7 trillion! — were all digital. Their rising share of online activity, with its hard-to-pindown impact on prices, profits and income distribution, has alarmed experts across the political spectrum.
This new and unwanted attention is not just about the economic impact of the giants. It is becoming increasingly clear that Americans (and their government) have little control over the digital platforms’ use of detailed customer data. And the 2016 and 2020 election campaigns made it dismayingly clear that social media’s capacity for spreading misleading information is a growing problem for democracy.
So, you’re no doubt now expecting this introduction to end with a demand for an aggressive new antitrust agenda aimed at curbing multiple threats to the economy and society. Such calls to arms are standard fare these days. But let me encourage you to pause and take a deep breath. I believe there’s lots to be done to make antitrust more effective. But a more cautious approach is warranted, lest antitrust become a scattershot substitute for focused remedies to offset some adverse social impacts of market forces.
In essence, modern antitrust is about preventing the creation or expansion of “market power,” the ability of firms to gain significant power over prices.
How We Got Here
Antitrust in the U.S. is primarily a federal matter, although all 50 states have antitrust laws as well, and many state attorneys general are becoming active where Washington chooses to remain passive. There are three main areas in which federal antitrust can bite:
Antitrust law bans collusive price-fixing among competing sellers as well as bid-rigging among competing suppliers. Agreements not to compete are in this category as well.
The law prohibits predatory behavior against actual and would-be rivals.
The law bars mergers and acquisitions where the consequence is likely to be a substantial reduction of competition.
In essence, modern antitrust is about preventing the creation or expansion of “market power,” the ability of firms to gain significant power over prices. Giant size itself, by the way, is not against the law; nor is modest size a defense for a firm caught colluding with rivals.
Notice, too, that modern antitrust policy has not been used as an all-purpose tool for dealing with all the ills of modern society. Indeed, over the past half century, the courts and regulators have seen it solely as the means to the end of increasing economic efficiency and the pace of innovation, which are generally thought to be byproducts of competition and beneficial to consumers.
Let’s briefly touch base on the three direct targets of antitrust.
Price-fixing. Prosecution of price-fixing is the least controversial element of antitrust. In the U.S., it is a criminal offense: individuals who collude can and sometimes do go to prison for it, while colluding firms may pay fines in the hundreds of millions. In addition, victims of price-fixing can sue for triple the damages they have suffered, and some judgments have resulted in billions of dollars in compensation. Various presidential administrations have had differing views on lots of things, but the pursuit of price-fixers enjoys widespread approval.
Mergers and acquisitions. The idea is to stop combinations that would probably lessen competition unless merger-related efficiencies would offset any deleterious impact on consumers. This means that antitrust enforcers must forecast the consequences of a merger — whether the merged firm would be able to increase prices, how this merger would affect costs, and whether market pressures would force the merged company to share the efficiency gains with customers.
This is a tall order, one that has led antitrust authorities to depend on data-driven statistical models. Reasonable people may well differ on how to interpret results and how to weigh those results against the inevitable claims that milk and honey will flow from the combination.
Monopolization. Notice that the operative word is not “monopoly.” If a company attains a dominant position in a market through superior skill, ingenuity or foresight — if, in essence, it builds a better mousetrap and the world beats a path to its door — antitrust enforcers have generally looked kindly on the outcome. To penalize the company for its success seems both unfair and a deterrent to further innovation.
It’s another story, however, if the company obtains or defends its dominant position by finding ways to exclude potential competitors. Among the behaviors that arouse suspicion: below-cost pricing for an extended period, an insistence that its customers or suppliers don’t patronize rivals, requirements that customers purchase a different item along with the one they really want.
As with mergers, reasonable people can disagree whether specific practices are benign or harmful (or a mix of both), and even whether a firm that has engaged in these practices possesses market power. No one would argue that a local auto mechanic’s insistence that all repairs be done with parts provided by their shop constitutes an antitrust violation. But what if a printer manufacturer insists that the ink and paper used in its machines be bought only from that company?
Wheels Within Wheels
Allegations of such transgressions only partly explain why the big platform companies like Google and Facebook are under suspicion. Indeed, only in the past few decades have economists begun to fit companies that don’t check the traditional boxes into the antitrust framework.
Network economics. In the world of traditional antitrust, value-creating relationships under analysis are typically bilateral — suppliers to producers, producers to customers, etc. Now consider a telephone system, where value to a customer turns on interactions with other customers. Adding another phone user generates benefits not only to the new user, but also to the others on the network who can now call or text him. This is called a “network externality,” or just a “network effect.”
Success early in the game can create outsize advantages that later contenders can’t overcome, especially if the in cumbent defends its turf aggressively and perhaps predates.
Network externalities have an important implication: there are usually advantages to large scale for both customers and the business. If individuals have a choice of networks, other things being equal, they will choose the larger network since the potential value will be greater.
This implies that these networks are likely to be characterized by “winner-take-most” outcomes, where smaller competing networks have difficulty surviving unless they offer some distinctive service not readily available from the dominant network. In any case, it will almost certainly be easier for smaller networks to compete if they can achieve interoperability with other networks.
Note another implication: the early stages of competition for a new market may well prove more important to success than laterstage competition. Success early in the game can create outsize advantages that later contenders can’t overcome, especially if the incumbent defends its turf aggressively and perhaps predates.
Multi-sided markets. Imagine another sort of market in which the value of the product or service to one distinct sort of customer depends on connected transactions with entirely different sorts of customers. There’s nothing truly new here. Consider a newspaper. One side of the market consists of the readers; the other, the advertisers. The value to advertisers increases with the newspaper’s circulation. Readers may or may not value the advertising. But the advertising helps pay for content, so the two sides are complementary. If a publication were to solicit content from third-party providers rather than relying on employees for content — the way scientific journals function — those content providers would make up a third side to the multi-sided market.
The newspaper also illustrates important points about pricing in multi-sided markets: the costs of serving the various sides will influence the price charged to each, but not necessarily in proportion to the cost of serving them. Some newspapers charge hefty subscription fees as well as charging advertisers; others distribute the paper free and rely on the increase in eyeballs to gain more revenue from advertisers.
Let’s be clear: the government has a legitimate role in market intervention to correct market failures, beyond just sustaining competition.
In this latter case, the zero price to readers is clearly below the direct costs of the paper, printing and distribution. This might be described as a “subsidy,” but it is better thought of as part of an integrated pricing structure that can’t be evaluated without reference to the other sides of the market.
What Antitrust Should Not Do
Like most aspects of life, even highly functional market economies are riddled with inefficiency and inequity. But that doesn’t mean antitrust offers the remedy. Here are some aspects of antitrust that, in my opinion, don’t need fixing.
The focus of modern antitrust is in maintaining competition. To widen the goals to include arguably worthy ends — such as expanded employment, a cleaner environment, improved income distribution and community development — may at times be tempting. But it would dilute and distract antitrust enforcement, since additional tradeoffs between efficiency and these other goals would need to be considered, and additional data and modeling would be needed to quantify these tradeoffs.
Let’s be clear: the government has a legiti mate role in market intervention to correct market failures, beyond just sustaining competition. But the current tasks of antitrust enforcers are already difficult; to broaden the objectives could make the tasks nearly impossible to do well.
Instead, if other issues loom, then complementary policies should be brought to bear. In the case of mergers that dislocate workers and communities, government has a legitimate role in helping the dislocated, a task that may include everything from job retraining to community development subsidies.
America’s long history of efforts to regulate competition with the goal of balancing the interests of suppliers, customers and communities has widely been acknowledged as a failure.
Proponents of widening antitrust to solve a host of social problems often understand they are wading in deep waters. But as a pragmatic matter, they see an opportunity to leverage merger approval to remedying assorted ills. What they often lose sight of is that “the other guys” will someday regain power, and then may use this opportunity to widen the goals of antitrust in directions that may not be so pleasing.
Don’t take my word on faith. America’s long history of efforts to regulate competition with the goal of balancing the interests of suppliers, customers and communities has widely been acknowledged as a failure. By the 1930s, all commercial interstate transportation — air, rail (freight and passenger), overthe- road trucking, buses — was regulated with respect to prices, routes, entry and exit. The same was true for telecommunications and for banking.
Competition suffered; entry was delayed or stifled; efficiency lagged. Deregulation of pricing and market entry starting in the 1970s, though far from problem-free, generally reduced prices and increased consumer options.
Similarly, restraint is appropriate in advocating the breakup of companies. Some advocates of expanded antitrust call for the dismemberment of large companies — especially the online giants. These advocates often point to the largely positive consequences of dismembering Ma Bell in 1984 and Standard Oil in 1911. Remedies on the table include restructuring Facebook, breaking it into multiple separate social networks and/or undoing the company’s acquisitions of Instagram (2011) and WhatsApp (2014).
Then there’s Amazon. Because Amazon is allegedly favoring its own branded products over those of third-party sellers, critics want to divide the behemoth into a marketplace platform (like eBay) and a separate company selling Amazon products. Similarly, Google should be separated into a pure search service and a second company providing all other services. As a model, these critics point to the 1933 Glass-Steagall Act, which forced the divorce of commercial banking (deposit-taking) from investment banking.
But history is an imperfect guide. First, the Standard Oil and AT&T breakups involved physical assets that could be readily separated. It’s unclear what breaking up Facebook would look like – perhaps just dividing its users among two or three individual social network clumps and hope they are sustainable? But if the network externalities are strong (as they probably are), the network that grows fastest will provide the most benefits to customers, and winner-take-most tendencies will prevail. Barring future intervention, Facebook would make a comeback (albeit with another name).
Perhaps the spinning off of Instagram and WhatsApp and their technologies would be feasible. But it’s unclear whether these freestanding entities would prosper today on their own, especially if Facebook competed with them with new apps created in-house. Note, too, that the spinoffs would make it difficult to achieve the sorts of efficiencies Facebook achieves on its integrated platform.
As for the Amazon split-up, it’s important to remember that big retailing chains of all sorts typically sell their own low-price “house brands” as well as national brands, and that these house-brand offerings are usually considered a pro-competitive feature. Amazon’s sale of everything from its own branded AA batteries to cargo shorts is similarly pro-competitive. If Amazon discriminates against individual firms seeking to use the Amazon platform, the problems should be addressed on a case-by-case basis with the antitrust tools already on the books.
As for the proposed Google split between search and browsing, the efficiency costs of de-integration are obvious. If a user wants a simple answer to “What time is it?”, a split that requires Google to direct the user to third-party sites would plainly inconvenience searchers.
An uncontroversial start would be to fund the two antitrust agencies adequately, so that they have the personnel and expertise to keep pace with the increasing size and complexity of the economy.
The issue on the table here can be summarized. When a major firm decides to integrate into adjacent activities, rivals in those activities are likely to feel the heat and complain. Deciding whether the integration is intended to put rivals at a disadvantage and reduce choice for customers or is just intended to increase profits by improving efficiency and serving customers better is always difficult.
The law, it’s worth remembering, already gives antitrust regulators the discretion to block mergers unless the merging companies sell key assets in order to sustain the premerger level of competition. This is more easily said than done, alas, because the regulators often have inadequate information about the relevant markets. Still, targeted asset divestiture has a better chance of sustaining competition than breaking up a company with a meat cleaver.
Sprucing Up Antitrust
While the current enthusiasm for regulation of highly concentrated markets seems misplaced, I do believe that some deft changes in antitrust could enhance competitiveness. An uncontroversial start would be to fund the two antitrust agencies adequately, so that they have the personnel and expertise to keep pace with the increasing size and complexity of the economy, especially in deciding the impact of mergers and acquisitions. But other changes would make sense, too.
Toughen merger enforcement. As John Kwoka, an economist at Northeastern, argued in the MI Review three years ago, antitrust enforcement with respect to mergers has grown lax. Too many mergers have been allowed that subsequently led to price increases. And there is no sign of change in this regard. The Department of Justice tried but failed in the courts to stop AT&T’s acquisition of Time Warner. And the agency waved through the merger of T-Mobile and Sprint with a remedy — encouraging the satellite TV provider Dish to enter the wireless market — that seems likely to fall far short of replacing Sprint as a viable fourth competitor. More vigorous enforcement would require a greater willingness to challenge proposed mergers in court and a sharper scalpel in shaping remedies that avoid court cases.
Study more of the “close-call” mergers that were allowed to proceed. Those who cannot learn from history are doomed to repeat it. Yes, I know, a hoary message, but apt nonetheless. If prices rise (in comparison with what otherwise would have been expected) after a “close-call” merger has been approved, then standards need to be tightened. If prices don’t rise, then the standards are in the right place or perhaps should even be loosened. We won’t know, though, unless we look.
Measure enforcement with a new ruler. Too many lawyers (and economists) judge the vigor of enforcement by counting merger challenges and monopolization prosecutions brought by the agencies. The problem with this approach is that it ignores the potential for deterrence. Consider price-fixing cases: if enforcement resources are plentiful, the penalties are high, and the criteria for what constitutes a violation are clear, one would expect few violations, and hence few agency cases. Thus, an increase in cases from one year to the next could be an indicator of greater enforcement effort, or of less because scofflaws take more chances.
Monopsony power, where a buyer has the ability to suppress the price it pays by reducing the quantity that it buys, is the full-fledge counterpart to monopoly.
Likewise with mergers. If the agencies’ criteria for challenging mergers were clear and consistently supported by the courts, then only a foolhardy CEO would propose a merger that would trigger a legal challenge. There would be few challenges, but only because few challengeable mergers would be proposed. The proper metric for measuring enforcement is to identify the boundary of challenge and non-challenge that has been set by the agencies and see how it squares with post-merger corporate behavior in terms of pricing and innovation.
Pay more attention to the market power of buyers. Almost all antitrust enforcement aims to limit market power on the selling side of markets, which is fine when the seller is Godzilla Industries and the buyers are you and me. But market power held by buyers – technically called “monopsony power” by economists – deserves more scrutiny.
Monopsony power, where a buyer has the ability to suppress the price it pays by reducing the quantity that it buys, is the full-fledged counterpart to monopoly. And the distortions it causes are parallel. Farmers – especially growers of perishables and livestock – have long complained about the market power of the middlemen who buy their output. Sometimes the complaints can be dismissed as backdoor pleas for more subsidies from Washington. But as the food processing industry has become more consolidated, these complaints may have gained greater validity.
Of parallel concern are “no poaching” agreements among employers (once a common abuse in Silicon Valley) who might otherwise bid against one another for technical specialists and similar “non-compete agreements” demanded by fast food franchises to prevent hamburger flippers from going elsewhere to flip for a higher wage. The latter case is especially egregious because the victims are often the working poor.
The enforcers already pay some attention; more is warranted.
Develop a consistent way to identify “relevant markets” in monopolization cases. The Justice Department developed a workable paradigm for identifying relevant markets – and thus market shares and market concentration – for merger cases some four decades ago. The model is used in almost all U.S. merger cases brought by Justice and by the FTC, and has been adopted by enforcement agencies abroad as well.
The same can’t be said for analyzing monopolization allegations. For one thing, the merger paradigm generally isn’t useful here because it applies to the prospective consequences of a combination, while monopolization involves allegations of already applied market power. The standard textbook treatment of monopoly stresses prices and profits that are higher than would occur under competition. But the competition yardstick often isn’t available for either prices or profits. And, in any event, the accounting vagaries that underlie profit calculations make profits an unreliable indicator of market power.
To be concrete: there is no rigorous way of determining whether Facebook currently exercises market power in advertising because there is no way of determining whether the relevant market is “all advertising on social network sites” (in which case Facebook surely has market power) or “all online advertising” (in which case Facebook may have market power) or “all advertising in all media markets” (in which case Facebook surely lacks market power).
Allow “indirect” purchasers to bring treble damages lawsuits. Since a Supreme Court decision in 1977, only direct purchasers are usually permitted to sue for treble damages in federal price-fixing or monopolization cases. If, say, a washing machine producer engages in price-fixing with other producers and sells the machines to distributors who then retail them to the general public, only the distributors — say, Best Buy or Lowe’s — can sue for treble the amount of the overcharge from the price-fixing. Consumers who bought the washers cannot.
The court reasoned that it would be too complicated to determine how much of the harm was suffered by the distributors and how much was passed through in the form of higher prices to their customers. I think the better data, better modeling and better statistical techniques now available make this decision worth revisiting.
More attention to state licensing and regulation of occupations. All 50 states license dozens of occupations. Sometimes licensing powers are exercised by government agencies, sometimes by semi-official boards of practitioners subject to little oversight by the state.
Although few of us would want to be treated by an unlicensed doctor, licensing of florists and hairdressers is harder to justify. Where public health and safety issues are remote, licensing serves primarily as a way to protect incumbents from competition.
Past Supreme Court decisions have granted the states immunity from the antitrust laws when imposing such restrictions, provided a state agency itself is doing the regulating or the state is actively overseeing the regulators. This leaves the federal antitrust agencies two potential points of leverage.
First, the FTC has challenged semiofficial boards when state oversight is weak and the boards engage in anticompetitive activity. For example, in 2015 the Supreme Court decided in the FTC’s favor in a challenge to the North Carolina Board of Dental Examiners’ efforts to bar spas and salons from selling teethwhitening kits.
It may well be that the network effects in large social networks are so strong that, absent intervention, the winner will never confront sustained competition.
Second, both the Justice Department and the FTC engage in “competition advocacy.” They testify at federal and state hearings on Supreme Court decisions have granted the states immunity from antitrust laws when imposing such restrictions, provided a state agency itself is doing the regulating or the state is actively overseeing the regulators. 24 The Milken Institute Review legislation that may have anticompetitive consequences, and they send comment letters on proposed federal and state regulations that may have anticompetitive consequences. Both agencies should step up these efforts, especially regarding licensing.
Expand the courts’ understanding of predatory practices. Since a 1993 Supreme Court decision, courts have insisted that plaintiffs prove the following in a predatory pricing case:
- That the accused firm had, or could hope to have, market power in the relevant market.
- That the firm was pricing its product below incremental cost for a sustained period.
- That the firm could reasonably expect to recoup its losses through the excessive profits it would earn after rivals abandoned the market.
Few plaintiffs have won their cases.
There are multiple problems with these criteria. First, they ignore the possibility that the predatory firm is willing to run losses simply to build a reputation for no-holdsbarred behavior. This discourages entry and encourages “get-along/go-along” behavior by its existing rivals. Second, the criteria are difficult to apply to non-price actions that could be predatory, such as insisting that its own distributors not handle the products of a rival. I think the antitrust agencies should try to educate the courts to the “reputation effect” that may well accompany predation.
Second, the agencies should encourage the courts to ask whether the expected profitability of the challenged action depended on the disappearance of the rival. This criterion has been called the “no business sense” rule. It would allow courts to more readily assess non-price restrictive actions that could be predatory.
Pare back statutory exemptions. Agriculture and fisheries are legally exempt from the reach of the antitrust laws; so is the insurance industry, as well as a few other industries. These weren’t good ideas when Congress enacted the exemptions, and they’re even less justified today. It’s long past time for repeal.
Where network effects are especially strong, consider interoperability as the remedy. It may well be that the network effects in large social networks are so strong that, absent intervention, the winner will never confront sustained competition. After all, even Google has tried to challenge Facebook’s lock on social networking, most recently in 2011, and has never succeeded in gaining a toehold. In any case, if network effects are strong, breaking up Facebook would undermine the value of the network to consumers — and would be likely to lead to a re-establishment of a dominant network within a few years.
But there may be a way around the consequences of market concentration: the feds could require Facebook to interconnect seamlessly with other social networks. This would make competitive entry plausible and, along with it, innovation.
There is precedent here. Federal regulations require telephone systems to interconnect, guaranteeing that all customers enjoy the network externality of being able to call pretty much anyone. Note, too, that when email systems developed in the 1990s, interconnection among the competing systems followed the same pattern.
An interoperability requirement for Facebook would mean that a user’s posts on Facebook would be automatically transmitted to designated friends on other social networks, and vice versa. It need not mean that Facebook would be required to share the information on users that it makes available to advertisers. But this posting interoperability could make entry by innovative startup social networks more likely.
Any interoperability requirement, with accompanying standards for interoperability, would require direct regulation with all the potential drawbacks. But as an alternative to a forced breakup, an interoperability requirement would be attractive.
Don’t let the complexity of multi-sided markets get in the way of constraining anticompetitive practices. The multi-sided nature of many network markets does indeed force analysts to set aside standard thinking on what constitutes anticompetitive practices. Nevertheless, if analysts can demonstrate the anticompetitive effects of a network’s restrictive practices, the burden of showing offsetting benefits should rest on the defendant network, not on the plaintiffs.
If a newspaper were to insist that its advertisers also place ads with the radio station the newspaper owns, or if the newspaper refuses to take the ads from an advertiser that has also placed ads with a rival newspaper, the burden for showing the socially beneficial effects of these practices should fall on the newspaper. The same logic should apply to the practices of online networks.
We’re Not Dead Yet
Antitrust policy is alive in the United States, but it is not as well as it could be. My armory of ideas for strengthening antitrust may not have the resonance of “Break ’em up!” But it would certainly satisfy the criterion of “First, do no harm.”