Big Tech looms large in the popular imagination because, in many ways, it is everywhere. Search engines provide us with instantaneous information. Social networks allow us to stay connected to friends and loved ones. Nearly any conceivable good or service is available within 2 days at the tap of a virtual button. To understand Big Tech’s importance in our lives, one need look no further than its centrality in keeping the economy limping along during the COVID pandemic.
But with this ubiquity comes a deep ambivalence toward the technologies that contribute so much to the fabric of modern life.
After 2016, the left blamed Big Tech for helping to elect Donald Trump president. In the last two years we’ve increasingly seen members of the right— like Senator Hawley—blame Big Tech for censoring conservative voices as “hate speech” or for other violations of the platforms’ terms of service. Other critics allege that as these firms grow large, their interests diverge from those of the U.S., and that the firms behave in a “treasonous” manner.
Although there are certainly important issues to discuss, these conversations have become increasingly muddied, leading some to regard tools like antitrust law as a policy cure-all for any behavior of a Big Tech firm that is found to be objectionable.
In a recent speech, Attorney General Barr described how the DOJ is taking a multi-pronged approach to rein in the perceived harms caused by Big Tech, including using antitrust laws. And there has been no shortage of (ex) presidential hopefuls declaring they will chasten Big Tech firms. Senator Warren had her plan to break them up, and intends to introduce a bill to the same effect. Senator Klobuchar introduced legislation that would alter antitrust laws in order to impose a legal presumption against “large” firms that would require them to defend their business conduct in court whenever a competitor complains that it has been harmed.
This view of antitrust law is misguided and, in the long run, threatens to undermine the rule of law.
Senator Klobuchar claims this bill “modernizes” antitrust but, in fact, it is based on an antiquated and long since abandoned concept of how competition laws should work. This approach to antitrust is rooted in the advocacy coming from a group of scholars and commentators often referred to as “neo-Brandeisian,” after former Supreme Court Justice Louis Brandeis, who was himself notoriously skeptical of large enterprises. Taking up Brandeis’s anti-bigness mantle, these critics charge that large firms by their very nature—especially Big Tech—are harmful to both the competitive and democratic processes. Further, these critics contend that modern antitrust law is incapable of adequately dealing with these harms owing to special features of Big Tech.
The neo-Brandeisians tend to frame their arguments in one of two ways. First, there is some ideal number of competitors in a given market that must exist before a market is competitive and, second, there are non-competition goals that enforcers and judges should consider when looking at antitrust cases (such as labor rights, political corruption, and income inequality).
Although they undoubtedly offer their criticisms in good faith, the neo-Brandeisians are misguided in their antitrust assault on Big Tech. Adopting these proposals would upend a century of legal and economic learning, and would ultimately fail to accomplish the advocates’ goals.
A Misguided View of Markets
The arguments against Big Tech are manifold, but there is a common underlying theme among them. Philosophical Jeffersonians, for instance, believe that industry concentration has grown to such an extent that society is facing a non-state form of central planning. Others allege that concentration has led to declining competition in the US, with concomitant reductions in the welfare of society. Others emphasize the need for a greater degree of fragmentation across industries in order to preserve liberty.
Underlying all of these complaints is a belief that when firms grow too large, they outstrip the ability of the law and market forces to discipline them. That is to say that in some sense “big is bad.” But, as modern economics has come to understand, it is not possible for a regulator to know in advance how many firms should exist and what size they should be. Indeed, in dynamic markets, like those in which Big Tech operates, it may not even be possible at all to ever reliably describe an optimal market structure.
Basing antitrust on firm size alone requires making an unsupported assertion that increases in concentration must result in a reduction in competition. But it could be that large firms willfully developed an anticompetitive monopoly, or it could be that the firms grew large by serving the demand of consumers and have simply outcompeted other firms. Without understanding more of the dynamics of a particular industry, superficially increased concentration can be just as consistent with good outcomes as with bad outcomes.
And in industries—like those in which Big Tech predominates—where significant up-front investment is necessary to reach critical mass, and where economies of scale are the norm, the effect of concentration on competition is even more ambiguous.
To take one example, as the US wireless industry has become more concentrated, prices have decreased. The price for mobile data has fallen from $49.00 per gigabit in 2010 to just over $6.00 per gigabit in 2017.
This is all to say: one cannot reliably maintain that “big is bad” as a mantra. Basing policy on a presumption about the size of firms will yield at best arbitrary results that fail to correct the alleged problems, and at worst, results guided by the biases of enforcers.
Network Effects Do Not Insulate Firms from Competition
More specifically, some neo-Brandeisian critiques contend that the peculiarities of how Big Tech firms work mean that traditional antitrust tolerance of large firms is no longer appropriate. This claim is rooted in the existence of network effects on many of the platforms, whereby the value of a particular service to its users is a function of the number and quality of other participants on the service. With tech platforms, these counterparties to a user can either exist on the same side of the network—e.g. friends on Facebook—or can exist on a different side—e.g. third-party sellers and buyers on Amazon.com.
One concern associated with network effects is that they lead to positive feedback loops that create winner-take-all markets. These markets are, it is assumed, then permanently insulated from competition by insurmountable barriers to entry.
This is a misguided view, particularly with respect to dynamic tech-driven markets. Despite the rapidity with which these services can grow large, these markets do not inevitably result in a single, unmovable winner. This is especially true where consumers have a variety of preferences, development of alternative goods and services is relatively easy, and multi-homing—for instance, installing and using both Snapchat and Twitter—is prevalent.
Crucially, even with network effects operating in a market with few competitors, there is no reason to believe that anticompetitive effects are certain to arise. There can be just as much competition “for the market” as there is within the market. This means that viable threats of entry from outside of a market can exert just as much discipline as direct competitors within a market.
Moreover, the existence of network effects do not in themselves prevent more innovative rivals from entering the market or from creating entirely new markets that compete for users’ attention, and the very network effects that led to the rise of a firm can quickly lead to its demise through a quickly multiplying “death spiral.”
Misunderstanding the Relevant Markets
Much of the misunderstanding from critics of Big Tech arises from a failure to properly understand how relevant product markets work. There are obvious cases—for instance the EU didn’t count Apple as a competitor to Google in its Android decision—but the problem is more subtle.
Facebook is a social network, Google is a search engine, and Amazon.com is an e-commerce platform. For Facebook and Google, their service is zero-priced from a user’s perspective. Facebook’s actual revenue driver is placing ads in front of its users. Google likewise earns its revenue by placing similar ads in front of its own users. Facebook and Google are, therefore, direct competitors in terms of how they derive revenue.
Amazon sells its own retail products, but also derives approximately half of its revenue from being, essentially, a product search service for its third-party merchants. Google directly competes with Amazon with its own product search service. In terms of gaining merchant participation on its platform, Amazon competes directly with Walmart and eBay. For the small amount of searches that are actually monetizable, Google faces numerous vertical competitors online and in mobile—for example, Yelp! is a vigorous competitor with Google in local search.
Stepping back, Amazon is often described as being dominant in e-commerce, accounting for roughly half of Internet sales. But the Internet is only a channel of distribution, not a market. It makes little more sense to brand Amazon as a retail monopolist for being dominant online than it would to describe JC Penny as a “department store monopolist.” The proper markets for evaluating the effects of Amazon’s retail business will be some share of retail, both online and offline.
When we look at firms like the Big Tech companies, what we are witnessing is the competition of ecosystems rather than narrow product markets. The recently announced acquisition of Fitbit by Google is emblematic of this point.
The chances that Google wants to get into Fitbit’s direct market are fairly low, as it’s been developing Wear OS for a number of years and hasn’t put the effort in to make it a top-tier wearables platform. But from a product market perspective this acquisition only fully makes sense when you consider the vector of competition across industry segments over the foreseeable future.
Not only are Big Tech companies trying to avoid their own Schumpetarian demise from unforeseen direct upstarts, they are competing vigorously with each other. The Big Tech firms all originated from different starting points, but are converging on a similar competitive struggle over devices and personal assistants. The merger presents a chance for more direct competition in this area, as part of the play is mobility and ecosystem development.
In short, competition, both actual and potential, exists and tempers the behavior of Big Tech. Not only do large tech firms constantly need to evaluate the viability of their own services to end-users in light of new entrants, these firms are constantly engaged in competitive struggles with each other. None of which even accounts for the large tech companies abroad, like Alibaba, Baidu, and Tencent, that would gladly enter the US market in force if a good opportunity arose to do so. See, for example, the ascendancy of TikTok.
Thinking about Big Tech in the preceding economic terms, however, is to some extent talking past the neo-Brandeisians rather than addressing their project directly. There is a non-technical sense of the word “monopoly” that the neo-Brandeisians employ in which a “monopoly” is just a dangerously large firm. But the development of antitrust law over the last one hundred and twenty years demonstrates that this view is underdeveloped and—knowingly or otherwise—advocates for this position wish to return the law to a primitive state.
The evolution of antitrust law in the US demonstrates that, to the extent that regulatory discretion and structural presumptions against “big” firms were considered, they were progressively rejected by courts, enforcers, and the academy.
The Evolution of Antitrust Law
At its founding, antitrust law was ambiguous in its aims. Its leading author, Senator Sherman, is oft cited for the declaration that “[i]f we would not submit to an emperor, we should not submit to an autocrat of trade[.]” Nonetheless the bill that became the Sherman Act was framed in terms that also resonate with contemporary antitrust theory and enforcement: the stated aim of the bill was to guarantee full and free competition, and to prevent consumers from facing increased costs.
Soon after passage of the Sherman Act, courts were forced to analyze these separate motivations for the new law. Despite the terse language of the Act, and its ambiguous reliance on economic concepts, courts began to develop a set of doctrines that were focused on separating acceptable conduct from unacceptable conduct. This was not an easy, task however, given the brevity of the law. Section 1 of the Sherman Act, for instance, declares that “[e]very contract… in restraint of trade… is declared to be illegal.”
Early cases took this prohibition literally. In 1897, the Supreme Court heard Trans-Missouri, where it had to consider whether the Sherman Act actually prohibited every such contract, or only contracts that were “unreasonable.” The Court surveyed the relevant historical record and, opting for a strict reading, declared the Act literally “renders illegal all agreements which are in restraint of trade or commerce.”
But, literally speaking, every contract is a form of a restraint of trade insofar as it binds two parties in a way unavailable to any other parties in the economy. In short order, courts began to recognize the problems with this interpretation. Just a year after Trans-Missouri Freight, the Sixth Circuit in Addyston Pipe carefully embroidered the harsh edges of the Supreme Court’s strict reading. In that case, Judge Taft recognized that the Sherman Act’s prohibition on every restraint of trade could not possibly be true as there were many restraints that were necessary in order to have a functioning economy. Judge Taft looked at the prior century of case law and noted many important restraints that the framers of the Act clearly could not have meant to make illegal. Thus he held it was only “naked” restraints—those designed explicitly to harm the competitive process and without any greater justification—that were illegal under the Sherman Act.
By 1911, the Supreme Court was ready to reconsider its strict interpretation. In Standard Oil, the Court was more solicitous of the earlier legal history, and—sidestepping its own previous refusal to focus antitrust law only on “unreasonable” restraints—acknowledged that “the freedom of the individual right to contract, when not unduly or improperly exercised, was the most efficient means for the prevention of monopoly.” In other words, some restraints on trade promote commerce, and some restraints hinder it. The goal of the antitrust laws, therefore, was only to prevent harmful restraints.
Thus, from this very early beginning, courts began to incorporate a more nuanced analysis, that, under a literal reading of the Act, could be viewed as illegal. None of this is to say that the evolution of antitrust doctrine was perfectly straightforward. As the twentieth century progressed, many cases exacerbated tensions between an economics-driven doctrine and a politics-driven doctrine.
Emblematic of this tension was ALCOA, which held that the antitrust laws existed both for economic reasons as well as to prevent industrial consolidation, even where such use of the antitrust laws may increase costs to individuals. Yet, the same opinion also noted instances in which the unique characteristics of a market meant that only one large firm could efficiently meet demand and handle the supply necessary for production. Thus, despite endorsing a Brandeisian interpretation of the antitrust laws, ALCOA acknowledged that “[a] single producer may be the survivor out of a group of active competitors, merely by virtue of his superior skill… although the result may expose the public to the evils of monopoly, the Act does not mean to condemn [such a result].“
ALCOA was not isolated in its difficulties coming to grips with the ambiguous commands of the antitrust laws. At times, holdings were based on fear that prices were too low and small competitors would have difficulty operating—in effect supporting price floors. In such cases, the Court found that under a protectionist reading of the laws it was acceptable that “occasional higher costs and prices might result from the maintenance of fragmented industries and markets.” At the same time courts continued to endorse price floors, however, they also continued to support a per se prohibition on price fixing, despite the economic similarity of the two. As Justice Stewart famously observed: “The sole consistency that I can find is that in litigation[…] the Government always wins”.
Despite these problems, as courts became more capable of economic analysis, doctrine evolved. By the late 1970s, the Harvard and Chicago Schools successfully introduced separate, but compatible, strains of thinking that disciplined enforcement. The outcome was the modern “consumer welfare standard,” a framework that requires courts and enforcers to evaluate both the costs and benefits of conduct, as well as the costs and benefits of proposed remedies.
In essence, antitrust doctrine evolved to focus on the welfare of consumers. Thus, the law cares about the competitive process and the welfare it can provide to society more generally, and ceased being solicitous of the misfortune faced by any particular firm, big or small.
Ultimately, this evolution of doctrine clarified that it is not the quantity or size of competitors that determines how healthy a market is on its own. Instead, different market conditions will be required to optimize consumer welfare in different contexts. Sometimes this is many sellers operating in an atomized market, sometimes this will be one or a small handful of firms providing most or all of the output.
Ambitious Social Policy Should Arise from Democratic Debate
None of this is to belittle the concerns of the neo-Brandeisians, but they fail to recognize the internal contradictions in what they seek.
For example one leading neo-Brandeisian, believes “the structure of our markets… can determine how much real liberty individuals experience in their daily lives.” This is because, “[m]ost people’s day-to-day experience of power comes… through relationships in their economic lives—negotiating pay with an employer… or wrangling the terms of business with a trading partner.” Consequently, she observes that Brandeis campaigned against large concentrations of economic power in a few hands, because “autocratic structures in the commercial sphere… threaten democracy in our civic sphere.”
Thus at least one animating factor for the neo-Brandeisians is concern about the effect of concentrated economic power over individual liberty: if large firms provide most of the jobs and produce most of the output, those firms have an incredible degree of control over society.
The contradiction arises from the fact that there is nothing special about large firms in this regard. In fact, if anything, small businesses with much tighter budgets and less operational leeway, have significantly larger incentives to deliver less than full value to consumers, and to demand more than they are entitled to from employees.
Further, in terms of production, large firms with efficient supply chains are better able to offer large quantities of quality goods at lower prices. The growth of relatively large firms in many sectors has led, on net, to dramatic improvements in living standards. When individuals enjoy a higher living standard—which is another way of saying that they are relatively less concerned about meeting their basic survival needs—they obtain far more agency in terms of shaping their lives as they see fit.
Large firms might be bad, but so too might small firms. Trying to make antitrust law focus on the size of a firm as a relevant metric gets the analysis backwards, and would ultimately fail to achieve the goals at which the neo-Brandeisians are aiming.
Further, if enforcers and courts were to roll back over a century of antitrust evolution in order to incorporate overtly political ends in enforcement, the net effect would be less democratic.
It is of course important to consider the broader questions society faces—the allocation of power among producers, workers and consumers; free expression; etc—but these are issues that should be confronted through a democratic process. Politicians answerable to voters are forced to confront these issues with care. Tailoring antitrust enforcement against Big Tech and other large firms commits a violation of liberty unto itself.
The tradeoffs involved in selecting one incommensurate social goal over another become lodged in the political organization of agencies, with firms and individuals no longer able to negotiate these tradeoffs amongst each other and through democratic elections. Further, by placing these issues in the hands of a small expert class, it inevitably invites rent-seeking.
The neo-Brandeisians’ campaign against Big Tech is an effort to return us to a bygone era before rigorous analysis was the norm. They want to make antitrust great again, as it were, in order to chasten Big Tech using theories that were retired in the twentieth century because they were found to increase political discretion without appreciably developing coherent competition law principles. It would be a mistake to join the anti-tech hysteria currently in vogue and reflexively upend a stable body of law that has helped nurture a dynamic market that benefits us all.