These new efforts to rein in big tech represent a confluence of many different factors: the rise of antitrust after decades of dormancy, the entrenched nature of big tech platforms in an industry that is used to the quick displacement of established monopolists by a new generation of upstarts, the rise of hate speech and misinformation online, growing fears about the loss of privacy, and instances of aggressive behavior by incumbents designed to disadvantage competitors, among others. Most of all, though, the focus on regulating or breaking up big tech is driven by the outsized influence these companies have come to have on almost all aspects of our lives. We spend much of our lives online—even more today in the wake of the pandemic. And most of us spend most of that online time using products and services provided by the four tech giants.
The desire to regulate the private actors that control so much of our lives is understandable, and some ideas for regulation make sense. But the political consensus in favor of regulating the tech industry is illusory. While everyone wants to regulate big tech, there is a wide range of (often contradictory) regulatory proposals on the table on issues ranging from misinformation to privacy to cybersecurity to copyright. These contradictions of platform regulation mean that it will be very hard to turn anti-tech popular sentiment into actual regulation, because the actual regulations some people want are anathema to others.
In the larger paper from which this is drawn, I discuss many such contradictory regulatory impulses. Here, I focus on just one: the array of antitrust claims against the tech giants. My argument is not that there are no antitrust concerns with tech platforms; indeed, I think there are. But the antitrust arguments commonly raised against those platforms frequently take contradictory positions. That requires us to choose among competing goals. This article identifies some of those contradictions. When we face hard policy choices, we should generally come down on the side of competition and interoperability that can open markets to new competitors rather than conduct-related regulation that entrenches incumbents and makes it harder for newcomers to compete.
Regulate What, Exactly?
These days it seems everyone wants tech companies to do (or not do) something, and they want government to require it. Fair enough; when any company has that large an influence over people’s lives, people have an interest in how they use that influence. But dig a little deeper, and the demands are not just diverse, but often flatly contradictory. While virtually everyone may want to regulate big tech, they don’t agree on what the government should require it to do (or forbid it from doing).
Antitrust, Advertising, and News Media.
The clearest example of contradictory regulatory impulses in antitrust enforcement centers around the collection and use of consumer data by platform intermediaries. There are three related regulatory impulses here, and they interplay in interesting ways. Those who would regulate platforms around data generally take one or more of the following positions:
- Platform intermediaries—specifically, Google and Facebook—are monopolies that are violating the antitrust laws by charging high prices in the advertising market, and with antitrust enforcement those prices will come down;
- Those platforms collect too much data about us and share it too widely with advertisers, and in a competitive market we would have more privacy; and
- Those platforms are stealing ad revenue from worthy sources like newspapers, and in a competitive market websites that display ads would generate more revenue.
Some propose privacy legislation to restrict access to data directly. But most of the regulatory proposals in this category focus on antitrust enforcement. The argument is that if we broke up the platforms, or restricted their market power in some other way, we would have a more competitive market, and reducing their concentrated control over our data would serve the substantive goals of increasing competition in advertising, protecting data privacy, and/or funding newspapers.
Notably, the U.S. government’s 2020 antitrust case against Google doesn’t take any of these approaches, though a coming state suit may. The Justice Department suit focuses on the fact that Google pays other companies to be the default search engine on various hardware platforms. That suit has its own problems—given that someone has to be the default and that the browsers are auctioning the position, it seems reasonable for Google to participate in that auction—but it doesn’t create the contradictions in the various regulatory proposals.
It’s hard to argue that at least some tech giants—Google in particular—don’t have market power vis-à-vis consumers. But the case for monopoly on the part of others is less clear-cut. Apple, Amazon, and Facebook may have no more than 60% of their respective consumer markets, and in Apple’s case considerably less. What Facebook and Apple do have are strong network effects, which lock in their customers and arguably justify treating their ecosystem as a separate market.
Even if they are monopolies, however, none of these companies charge the high consumer prices we expect from traditional monopolies. Indeed, Google and Facebook provide their consumer services for free, and the antitrust objections to Amazon have centered on the argument that its prices are too low. As a result, many antitrust complaints have focused not on the consumer market but on the place platforms make most of their money—advertising. They argue that Google and Facebook (and sometimes others as well, though it’s a bit awkward to point to multiple “monopolists” in a single market) dominate the market for online advertising. They do this because they collect data about their users—the price of free Internet goods —and use that data to target more effective ads to consumers. The most common objection is that this invades our privacy. But others—notably news outlets—also object that the Google-Facebook dominance of digital advertising leaves no room for them to make money from advertising of their own. And all this is styled as an antitrust problem—it is the dominance Google and Facebook exert over advertising that leads them to collect too much data and drive out competitors who also want to host ads.
The problem is that that theory makes no sense as an antitrust matter. Google and Facebook are selling advertising space. If they dominate that market —whether because their access to consumer data gives them an edge in targeting the ad space or because they reach so many people they are viewed as must-place destinations for advertisers—they should be able to charge advertisers a premium for that space above a competitive market price for advertising space. That might be a bad thing if you’re an advertiser: you’re paying more money than you would in a competitive advertising market. But Google and Facebook are not the only companies serving Internet ads, and those higher prices would drive advertisers (and sites that host advertising) to competitors unless Google and Facebook are significantly better at targeting those ads than competitors are.
But even if they are dominating the market and increasing prices for digital advertising, consumers may be better off as a result of the conduct that is the source of many of the complaints about Google’s dominance of digital advertising. If Google and/or Facebook are advertising space monopolists, they are charging more money for ad space than they would in a competitive market and offering fewer ads as a result. That means they are using your consumer data to sell ads to fewer people at a higher price than they and others would if the advertising market were competitive. Advertisers might not like that because it could reduce the effectiveness of their ads, but most consumers would probably view this as a benefit. Intervening to bring competition to digital advertising—whether by breaking up the platform giants or opening up their data troves to competitors—means that more companies will compete to sell more of your personal data to advertisers for less money. Indeed, the Texas attorney general, who recently filed an antitrust complaint against Google, has argued that Google should be forced to share its data about users with rivals. That might well be a good thing from an antitrust perspective, since it opens the market to competition, but it is the opposite of what most people complaining about platform dominance of data and advertising actually want.
Similarly, news organizations compete with Google and Facebook to sell advertising space to those who come to their websites. If Google and Facebook are charging more than the competitive rate for advertising space, that presents an opportunity for others—like news organizations—who compete with them to sell advertising space. Either they can charge a competitive price and take sales away from their overpriced “monopoly” competitors or they can shield under the Google-Facebook price umbrella and charge more for their advertising space. In either case, a competitive market is the last thing they want. It will reduce their advertising revenue further.
There may be good reason to think Google and Facebook dominate digital advertising, raising prices to advertisers and reducing the output of ads based on personal data. There may be good reason to think Google and Facebook use too much of our data too cheaply to target ads. And it may be possible that Google and Facebook are unfairly taking ad revenue away from news outlets that depend on it. But it is not possible to coherently think all those things at once. And regulation designed to fix one of those perceived problems—say, bringing more competition to digital advertising—is likely to make the other problems worse.
None of this is to say that there isn’t an antitrust problem with big tech; indeed, I suggest below that there is. Nor is it to suggest that there is no problem with their control over digital advertising. If the allegations that various states have made that Google and Facebook have secretly colluded to prevent competition, or that Google manipulates advertising auction results to disadvantage competitors, turn out to be true, those are problems the law should address. But antitrust is not capable of giving everyone everything they seem to want from it, because some of what they want is self-contradictory.
Breaking up data monopolies.
A similar problem affects the widespread concern about data monopolies. Privacy advocates worry that dominant Internet platforms know too much about us, aggregating data from a variety of sources to build an accurate profile of our needs and desires. But this is actually two different worries coupled together. Many worry that too much information about us is readily available on the Internet. They seek to reduce the flow of personal information on the Internet generally. Others focus on the concentration of personal information into a few hands, giving dominant platforms an information edge no one else can match.
While both concerns focus on consumer privacy, they have different, even contradictory, solutions. Those concerned about the widespread flow of personal information generally want to make it harder for companies to sell or otherwise share that information. Restrictions on sale of personal data or cookies that track people across websites can reduce the flow of information among companies. But because it is easier to regulate the flow of information than to force companies to “unlearn” information they already have, these proposals tend to reinforce the dominance of existing Internet platforms by making sure that they will always have an edge over potential new competitors. For those concerned with the concentration of private information in the hands of a few companies, these changes may make the world worse, not better.
Conversely, those concerned with concentration of data on the Internet may want to break up the big data monopolies, if not the companies themselves. Doing so may reduce the dominance of incumbent Internet platforms, but it is likely to do so by broadening, not restricting, the flow of private information. That is most clearly true of efforts to break up the Internet monopolies outright. But even efforts to combat the control those dominant firms have over personal data involve making sure that they don’t end up with more access than their competitors, for instance by regulating exclusive deals among firms or by allowing user data portability or efforts to scrape sites to build interoperable ones. Each of these approaches, however, is likely to involve more personal data in the hands of more companies, since new entrants will be unable to compete effectively against incumbents without access to the same data those incumbents have. That may be a good thing, but it is unlikely to be what privacy advocates want. Indeed, Google had to put off its Privacy Sandbox initiative to phase out third-party cookies—something privacy advocates would like—because UK antitrust authorities thought that interfering with third-party cookies could cement Google’s market dominance. Antitrust may well suggest that it would be desirable to break up data monopolies, but doing that is likely to produce the opposite effects of what privacy advocates complaining about those monopolies actually want.
Navigating the Contradictions of Tech Antitrust
What are we to make of these fundamental contradictions in basic efforts to regulate the Internet? Are contradictions inherent in the very idea of regulating the Internet? Does it mean we should give up the idea of regulation altogether?
I wouldn’t go that far. But I do think these contradictions should make us wary about rushing into regulation. Real regulation of technology platforms is likely to require difficult tradeoffs, giving some people what they want but making things worse in other respects. Sometimes the right response to that is to forebear from regulation altogether. Other times regulation may be necessary, but we need to go into it with eyes open, understanding the harm it will likely cause. Finally, the nature of these tradeoffs, and the demand for contradictory things from platform regulation, may ironically point the way towards the kind of regulation we do need. That is particularly true when the regulations do not just create tensions, but affirmatively demand that Internet platforms act in opposite ways.
Not all contradictory platform regulation proposals warrant rejecting regulation altogether. Sometimes the contradictions in platform regulation proposals reflect not the futility or absurdity of the idea of regulation, but the fact that regulating complex industries like online intermediaries involves tough policy tradeoffs. Sometimes there are good policy arguments on both sides. But the contradictions of these regulatory proposals mean that we can’t just say “regulate” and get the best of all possible worlds. We need to choose one benefit at the expense of the other, or balance the two, accepting, for instance, less privacy in order to have more competition or vice versa.
But compromise may not always be possible, and antitrust regulators will have to choose among competing policy goals. I think this is likely true of data monopolies and the advertising market. There are reasons to think those markets are being monopolized, resulting in a less competitive advertising market and higher prices to advertisers, and giving incumbents a structural advantage over those who lack access to users’ personal data. That seems to be something we might want to correct. But doing so would make other problems worse. Advertisers could place ads more cheaply in a competitive market, but that would probably mean passing less money on to news organizations and others who depend on that revenue (and who have been behind things like Europe’s newspaper link tax). To be fair, a competitive market might reduce the intermediary charge a Google or Facebook can take when placing ads on third-party sites, potentially allowing cheaper advertising rates without reducing how much gets passed on to the sites where the advertising is placed. On the other hand, in a competitive intermediary placement market each intermediary will likely have less information about customers with which to effectively target ads, reducing its efficacy and therefore the revenue it would generate. Predicting which of these effects would predominate is hard.
Breaking up data monopolies presents a clearer choice. Today we have a few dominant companies who know us intimately and can deliver us the information (and ads) we want. But we are beholden to them. We could solve the dominance problem by mandating the sharing of our data, perhaps introducing effective competition in the platform market but also reducing our privacy significantly by allowing many different companies to know everything about us. Allowing what some have called “adversarial interoperability” would permit the development of effective competitors to companies like Facebook, for instance, by requiring it not to close off its APIs to potential competitors, allowing consumers to port their data to other platforms. But doing so would mean less privacy, as more companies obtained our data and competed to use and sell.
Alternatively, we could prevent incumbents from selling or sharing our data with third parties, protecting what privacy we have left but ensuring that the incumbents have a major advantage over any challengers. Favoring privacy may mean less competition, locking in the incumbent tech companies indefinitely.
Or we could—perhaps—prevent the incumbents from using the information they already have about us, enhancing privacy and perhaps even leveling the playing field, but at the cost of making their products far less useful to us (and, as just noted, reducing revenues to sites like news organizations that depend on their targeted advertising platforms). I have preferences among these options, some of which I discuss in the next Part; others might have different preferences. But they all involve difficult choices, and we can’t all get what we want.
Confronting the Outsized Footprint of Tech Giants
I think there are lessons to be learned from thinking about why there are so many contradictory efforts to regulate big tech. The answer has to do with the outsized role these companies play in our lives. They are private companies, but for most people their decisions may as well be government mandates, given the lack of control we have over them. Yes, you can quit Facebook and Instagram, or decide to use DuckDuckGo rather than Google for all your searches. Maybe you can avoid both Apple and Google, though that probably means you won’t have a smartphone. But making those choices means making fundamental alterations to your lifestyle. The dominant tech firms are private companies, but they have been so dominant for so long that they feel more like institutionalized monopolies, the kind we used to regulate as public franchises. And so people who don’t like the way they do something naturally react by thinking not “I’ll switch to an available alternative in a competitive market” but “someone should do something, and the government is the only one who can.” And that means that whatever people want from tech giants, they increasingly turn to regulation to get it.
That’s not good. We treat private companies differently than governments precisely because we think markets give consumers choice and the risk of consumer exit gives companies an incentive to respond to consumer demands. When that isn’t true for some structural reason, as with natural monopolies, we generally declare the company a public franchise, give it exclusive rights to provide a particular service, and subject it to price and quality regulation. That seems to be the rationale for many, though not all, of the proposed regulations of the tech industry.
But public franchise regulation should be a last resort, for a number of reasons. First, governments aren’t very good at promoting efficient and cheap products, and cost-plus regulation often leads to bloat in regulated industries. Further, regulated industries tend to rest on their laurels, insulated from competitive threats and unlikely to reap extra profits for innovating. Regulated industries from telephones to electric power to taxicabs spent decades making few significant improvements to their products until they were deregulated or faced outside competitive threats. Regulations are also subject to industry capture, with the regulators benefiting the very industries they are supposed to scrutinize. And once government creates a comprehensive set of regulations for an industry, it makes it harder for others to break into that industry, since they don’t have the experience dealing with those complex regulations.
Many of the industries we viewed as natural monopolies in a prior era (including telephones, airlines, railroads, and even electric power generation) turned out to be potentially competitive markets. But regulation tends to perpetuate itself even when competition becomes feasible. And regulation imposes significant costs. That’s not to say there is no room for traditional public franchise regulation. There is, just as there are services (education, police, etc.) best provided by the government rather than private parties. But we should treat an industry as a public franchise only if we really have no other choice.
A better alternative is to try to inject real competition into a tech industry that has been lacking it in recent years. Andrew McCreary and I have made one suggestion along these lines: Make it harder for incumbents to buy startups that might otherwise grow into competitive threats. But if the goal is to open tech markets to competition so that competition can obviate the need for much proposed regulation, more will be required, at least in the short term.
This suggests that when we face hard policy choices, we should generally come down on the side of competition and interoperability that can open markets to new competitors rather than conduct-related regulation that entrenches incumbents and makes it harder for newcomers to compete. That means favoring antitrust enforcement that demands structural separation, or at least imposes nondiscrimination rules on self-dealing by vertically integrated monopolists, over efforts to lock in data monopolies to protect privacy or mandate supracompetitive payments to newspapers and the like. It means favoring interoperability at the expense of privacy, turning away contract and Computer Fraud and Abuse Acts claims that dominant sites have used to prevent third parties from offering products or services that interconnect with dominant firm sites and perhaps using antitrust or other tools to force interoperability. And it means preserving laws that give tech companies the freedom to decide what content to allow on their site over alternatives that either mandate detailed scrutiny of content or forbid that scrutiny and treat tech platforms like government actors. Those laws protect content decisions, but they arguably should not extend to anticompetitive acts designed to block a competitor from interconnection.
These are, as I noted above, hard choices. Privacy advocates, newspapers, and both the people who want more content removals on the Internet and those who want less all have reasons for their policy preferences. And there may be ways to achieve some of those goals (like strengthening consumer privacy rights or funding local newspapers) directly, without using the regulation of tech companies to achieve that goal. But in the long run, regulatory choices that impose obligations on incumbents to do the things we want them to do as a matter of social policy are likely to entrench those incumbents, making it harder and more costly for someone to compete with them and eliminating the possibility of competing by offering a different set of policies. Effective antitrust enforcement that opens tech markets to competition may, by virtue of that competition, get people many of the things they seek today from regulation—a choice of Internet platforms with more privacy versus more free services, more content restrictions versus less, and other changes. It may also reduce the perceived need for regulation because consumers will have more choice and therefore more voice. The reverse will not be true. If we lock in detailed policy requirements for privacy, content regulation, linking to newspapers, and many more things, we are likely resigning ourselves to a long reign of continued dominance by the current incumbents, and a future of governments deciding what those dominant firms can and can’t offer. That is something we should avoid if we can. Conversely, a warning to tech companies: if we can’t promote effective competition through interoperability, the alternative may be detailed, often contradictory regulation.
Conclusion
Tech platforms have an enormous influence on our lives, and they are not materially constrained by competition today. Those facts have led to calls for their regulation from across the political spectrum. Many of those calls demand contradictory things, a sign that we increasingly want tech regulation (and tech platforms) to be all things to all people. Regulation can’t serve that need. At a minimum, regulating the tech industry will require hard choices between competing goals.
But it may also suggest that our instinct that regulation will solve the problems with the tech industry is misguided. Regulation seems like the only option we have because competition hasn’t been a realistic option in the tech industry of late. It is that latter problem we need to solve. A focus on reintroducing competition into tech platforms will obviate the need for some regulation altogether, and it should guide us in deciding how to make the tough choices that tech platform regulation entails.
