Debunking the "Big is Bad" Bogeyman: Facebook Benefits Consumers

Category: Print Issues

Debunking the "Big is Bad" Bogeyman: How Facebook Benefits Consumers

Christopher Marchese

Volume 28

, Issue 1

28 Geo. Mason L. Rev. 1 (2020)

Introduction

“Ever since Congress overwhelmingly passed and President Benjamin Harrison signed the Sherman Act in 1890,” recounts the Supreme Court, “‘protecting consumers from monopoly prices’ has been ‘the central concern of antitrust.’”¹

Not exactly. Although that more or less describes today’s understanding of Congress’s landmark legislation, it sweeps decades of muddled antitrust thinking under the rug. The truth is, the federal government and the Supreme Court struggled to apply Congress’s broadly worded antitrust statute for almost eighty years. Neither had a sound method for sorting procompetitive business actions from anticompetitive ones. So they relied instead on gut instinct: that big is bad.² As it turns out, that I-know-it-when-I-see-it approach led to scattershot enforcement decisions, higher prices, and lower output.³

All that changed in the 1970s. Led primarily by the Chicago School of Economics and Robert Bork, economists persuaded the government, courts, scholars, and practitioners that antitrust laws should be rooted in empirical evidence.⁴ After some finetuning, that idea became a bipartisan consensus: antitrust laws should protect consumers and the benefits they receive from a competitive marketplace—lower prices, higher-quality goods, innovative products, increased productivity, and so on.⁵

Known as the “consumer welfare standard,” this understanding still enjoys widespread support today.⁶ But that support is not universal. As the United States grapples with growing concerns over headline issues like income inequality, some have called for a new approach to antitrust enforcement.⁷

Except the approach isn’t new. It’s a return to the antitrust doctrines of yesteryear—when the government thought big is bad, small is good, and inefficiency and higher prices are sometimes worth it (“it” often left undefined).⁸ To be sure, some scholars would tweak the old approach here and there.⁹ At bottom, though, they’d have the government return to using the blunt instrument of antitrust enforcement to pursue policy goals beyond protecting consumers. This approach has spilled over into politics, too, catching the eye of the contemporary left and the populist right.¹⁰

But this new-old approach is still an outlier. So much so that it’s been derisively dubbed “Hipster Antitrust.”¹¹ (More charitably, it might be called “Neo-Brandeisian Antitrust,” after Supreme Court Justice Louis Brandeis, who believed big really was bad.¹²) Whatever the name, it comes with considerable baggage: decades’ worth of evidence and experience proving that the consumer welfare standard works remarkably well.

Savvy sympathizers of the movement realize this. Rather than call for full-scale reform, they market their ideas as mere “updates” to antitrust doctrine.¹³ Even savvier scholars claim their desired policy goals are achievable under existing doctrine—no changes necessary.¹⁴ But with just a little digging, their shallow treatment of case law and shoehorning of evidence gives it all away: their arguments rest on substituting “social welfare” (as they envision it) for “consumer welfare” and “subjective” for “standard.”¹⁵

Such is the case in a series of recent papers written by Professor Fiona M. Scott Morton and David C. Dinielli.¹⁶ The duo has so far published three under the title “Roadmap for an Antitrust Case Against . . .” Google, Facebook, and Google again.¹⁷ As I explained elsewhere in a response to their first paper against Google, there is no antitrust case against either company.¹⁸

This Article, though, focuses solely on debunking Scott Morton and Dinielli’s claims that Facebook is a monopoly and that it harms users, advertisers, and publishers. As this Article shows, Facebook is popular, but it’s no monopoly. And it shows that consumers benefit immensely from Facebook.

I. Setting the Scene

A. The Social Network

In 2004, Mark Zuckerberg and other Harvard students launched TheFacebook.¹⁹ The platform initially allowed only Harvard students to join.²⁰ But within a year, TheFacebook dropped the The and opened to students from other colleges.²¹ By 2006, Facebook was open to anyone thirteen or older with an email address.²² Today, roughly two-thirds of Americans have joined the site, as have 2 billion people worldwide.²³

Safe to say, Facebook is the social network—at least for now. Even as the platform attracts billions of users, adapts to changing consumer tastes, innovates and improves, Facebook faces fierce competition.²⁴ It is neither as dominant as its critics claim, nor as safe as it would prefer.²⁵ The platform that has attracted nearly seventy percent of Americans has also seen younger Americans jump to competitors.²⁶ The platform that has made billions in advertising revenue has also seen those advertisers boycott it.²⁷ And the platform that was once heralded as an American success story has also been raked over the coals by American politicians.²⁸

Facebook’s founding motto was “move fast and break things.”²⁹ That it did. The platform quickly expanded from the elite halls of Harvard to the American mainstream.³⁰ In doing so, it broke MySpace’s dominance in social networking—a feat once thought unimaginable.³¹ And it broke the comfortable channels of news, commentary, and thought that traditional media, and the political class reliant on it, controlled.

Facebook gave all Americans a stage, a microphone, and an audience, all the while connecting those Americans to their friends and family, matching consumers with businesses large and small, and mixing silly with serious, professional with personal. It became, in other words, whatever each of us wanted it to be.

With the decades-old status quo extinguished, Facebook forever changed how we interact online. But despite all this breaking—of limits, expectations, possibilities—Facebook could not break one truism in technology: innovation and the competition it unleashes are always lurking.

Facebook faces pressure from outside the marketplace, too. Like its competitors, Facebook finds itself in the crosshairs of this decade’s culture wars.³² The left criticizes the platform for not removing enough content that it believes is dangerous to American democracy;³³ the right says the opposite, that it removes too much content and that any such censorship undermines American values like free speech.³⁴ Implicit in both criticisms is a recognition of Facebook as an institutional force that needs to be reckoned with, lest it (i.e., its users) continue to compete with society’s gatekeepers (e.g., traditional media, the political class, elites).

This recognition has led some on the left and right to rediscover an old tool: use of antitrust to pursue political ends.³⁵ That some politicians, bureaucrats, and scholars wish to use the blunt force of antitrust to achieve their desired ends is unfortunate—but not new.³⁶

B. Scott Morton, Dinielli, & Their Roadmap

Scott Morton and Dinielli have now penned three “Roadmaps,” each of which relies on flawed preliminary analysis conducted by the United Kingdom’s Competition and Markets Authority (“CMA”).³⁷ The CMA has since released its final report on competition in digital advertising, concluding that Facebook and Google dominate separate digital advertising markets and that their dominance harms consumers, including users and advertisers.³⁸ Relying on the report, the CMA has pleaded with the UK government for more power, including the ability to break up industries without having to abide by existing protections that ensure such an upheaval in the marketplace is justified.³⁹

Even setting aside the fact that the CMA’s data and analysis are not focused on the United States, neither Scott Morton and Dinielli nor any government should rely on the CMA’s advertising reports—preliminary, final, or otherwise. Despite spanning hundreds of pages, the CMA’s reports are built on a foundation of sand.

For starters, the CMA report is untethered from actual antitrust doctrine. First, the CMA failed to uncover any actual harms. Instead, it merely speculates about how consumers “might” be worse off because of Big Tech.⁴⁰ Crucially, the report does not find that consumers actually are worse off. No wonder the CMA couches its “findings”—more accurately called hypotheses—in wishy-washy language: “could be” appears 91 times; “we believe” 27 times; and “might” 76 times.⁴¹ Indeed, those phrases mirror what one might expect to see in a press release announcing the start of an investigation, not in a final report after a year-long process concludes. This is especially true given the stakes: breaking up some of the most successful companies in history.

Second, the report snubs antitrust analysis. To be sure, the CMA sees things differently. It contends that advertising prices are higher than they ought to be, which it takes as evidence of anticompetitive markets.⁴² But the CMA fails to explain how it knows this. It doesn’t even bother to give a ballpark number on what prices would be in the hypothetical competitive market it constructed for the analysis it does not share with readers.⁴³

At the same time, the CMA dismisses most of the benefits Google and Facebook have brought to the digital advertising market. There is no mention of price reductions—which, in the United States amount to over forty percent since 2010.⁴⁴ It also gives no consideration to targeted advertising’s benefits, which allow small- and medium-sized businesses to reach consumers across the country.⁴⁵ For less money than ever before, these businesses no longer must rely on their local markets for consumers; now, they can reach consumers everywhere.⁴⁶

The CMA’s report and Scott Morton and Dinielli’s reports give shallow treatment to actual evidence. They ignore or downplay evidence of benefits to consumers while hyping speculative evidence of harm that accords more with their preferences than with consumers’. Their arguments boil down to: we know what consumer harms look like, we don’t see definitive evidence of those harms, but we have a hunch such harms exist, so we will speculate about those harms, and cast them as conclusions based on facts.

This Article, then, should be read as a rebuke of both Scott Morton and Dinielli’s Roadmap and the CMA’s final report. While they are scholars, not government officials, and while the CMA’s jurisdiction does not extend to the United States, their ideas are so far astray—so unglued from antitrust doctrine—that they must be called out for what they are: a return to “big is bad” no matter the evidence.

II. Primer on Antitrust Laws & Doctrines

Before jumping headfirst into the Roadmap’s analytical defects, it’s worth getting a feel for antitrust law and its doctrines. Although this Section will sound familiar to antitrust aficionados, much of its content is conspicuously missing from the Roadmap’s narrative.

A. Overview

Congress passed the Sherman Antitrust Act in 1890.⁴⁷ And it did so in response to mounting unrest brought on by the country’s rapid transformation into an industrial giant after the Civil War.⁴⁸ This period witnessed factories displace farms, cities overshadow towns, and corporate America edge out local business.⁴⁹ But it was the great trusts of the day—the Standard Oils—that cast the darkest shadow over the country.⁵⁰ Their seemingly endless combinations and consolidations of industries stirred Congress to enshrine in law the country’s “national values of free enterprise and economic competition.”⁵¹

But, like other national values, free enterprise and economic competition invited disagreement among Americans.⁵² It also bedeviled the US Supreme Court. In an early case, the Court severely narrowed the Sherman Act’s reach by exempting manufacturers, rendering most of the economy outside the bounds of the government’s enforcement power.⁵³ That controversial decision soon gave way to an antitrust doctrine that bobbed and weaved from one social concern to some other economic concern and back again. Along this zig-zagging path, the federal government was able to:

* “[P]revent ‘bigness,’ that is, to preserve the small, localized businesses that characterized early America”;⁵⁴

* Protect “small dealers and worthy men” from their competitors even if doing so raised prices for consumers;⁵⁵

* Block “great industrial consolidations [because they] are inherently undesirable, regardless of their economic results”;⁵⁶ and

* Shield “viable, small, locally owned business” from competition even if that meant “higher costs and prices might result from the maintenance of fragmented industries and markets.”⁵⁷

Ripped from any objective standard, these antitrust cases transmogrified a law meant to champion competition and protect consumer welfare into doctrines that shielded the status quo and handed out corporate welfare.

These were bad days for antitrust. As former Federal Trade Commissioner Joshua Wright put it, “[t]he result of this approach was that consumers were made worse off by preventing the very competition from which they would benefit and which the competition laws were supposed to promote.”⁵⁸ “In the name of defending helpless individuals,” he adds, “the Court decreased the purchasing power of individual consumers—by preserving inefficient firms with higher prices and lower output—and issued opinions that explicitly chose to foster corporate welfare over consumer welfare.”⁵⁹

Another problem: No one knew what conduct was permissible and what wasn’t. Without clear, objective standards for the government to apply equally, regardless of a business’s size, the business community was left to guess—and to litigate should the government decide that particular business was that year’s antitrust target.⁶⁰ As Justice Potter Stewart of the Supreme Court once remarked: “The sole consistency that I can find is that in litigation under § 7 [of a related antitrust act], the Government always wins.”⁶¹

Once the consumer welfare standard entered the scene, however, antitrust improved. Today, antitrust focuses on promoting the benefits that come from national values like free enterprise and market competition⁶²: lower prices for consumers, lower costs for businesses, higher productivity for both, and greater innovation across the board.⁶³

And today, we know that antitrust laws are not meant to protect a business’s competitors but instead the competitive process. “The purpose of the [Sherman Antitrust] Act,” the Supreme Court has said, “is not to protect businesses from the working of the market; it is to protect the public from the failure of the market.”⁶⁴ And because “[c]ompetition is a ruthless process,” US antitrust doctrine tolerates aggressive conduct even when “[a] firm that reduces cost and expands sales injures rivals—sometimes fatally.”⁶⁵ In fact, even when a firm has monopoly power—what most people think of when they hear antitrust—the law applies only when that firm takes illegal steps to entrench its monopoly power.⁶⁶ In other words, monopolies are not, by themselves, illegal.

B. Sherman Antitrust Act § 2: Monopolies

Against this backdrop, this Article turns next to the Sherman Act’s text. Under section 2 of that law, no business may “monopolize, or attempt to monopolize, or combine or conspire with any other person or persons, to monopolize any part of the trade or commerce among the several States, or with foreign nations.”⁶⁷ A business violates the law’s ban on monopolization only when it:

(1) Has durable monopoly power; and

(2) Uses exclusionary practices to obtain, maintain, or increase its monopoly power.⁶⁸

1. Step 1: Show Monopoly Power

Monopoly power means a business has “the power to control prices or exclude competition” from the market.⁶⁹ Plaintiffs can show monopoly power through direct evidence that the defendant business charged prices significantly higher than what they would be if the relevant market were competitive.⁷⁰ More often, though, they rely on indirect evidence that shows a business:

(1) Has a large share of the relevant market it operates in; and

(2) Is protected by barriers to entry into that market.⁷¹

Monopoly power requires “something greater” than just mere market power.⁷² In other words, the defendant business’s market share must be relatively high. Plaintiffs thus have an incentive to define the market narrowly (fewer competitors, more dominance); defendants, the opposite (more competitors, less dominance). The Supreme Court has sought to avoid these subjective pitfalls by defining the relevant market as including the product or service at issue and anything “reasonably interchangeable” with it.⁷³ Whether a product or service is reasonably interchangeable turns on whether a price increase for Product A would lead consumers to buy Product B instead.⁷⁴

Although the Supreme Court’s definition dates to 1962, the Department of Justice and the Federal Trade Commission—the plaintiffs in many monopoly cases—also use demand substitution to define the relevant market.⁷⁵ Their “hypothetical monopolist” test asks whether a monopolist would profit from imposing a “small but significant and non-transitory increase in price” (“SSNIP”) of about five percent on the product in question.⁷⁶ If it would, then that is the relevant market; if not, then the market is expanded until buyers have no substitutes left.⁷⁷

But there’s a wrinkle in this case: Facebook is free. Although some lower courts have held that free products or services are exempt from antitrust enforcement,⁷⁸ some scholars⁷⁹ have argued that nonprice factors—for instance, personal data shared—can substitute for price. For purposes of this Article, the latter approach will be taken.

In any event, there is no magic number that declares dominance, but the Supreme Court has never found monopoly power when the market share is below seventy-five percent.⁸⁰ Some circuit courts have tried to flesh this analysis out with rebuttable presumptions: in the Tenth Circuit, the market share must usually be at least seventy percent;⁸¹ in the Second Circuit, usually ninety percent;⁸² and in the Third, “significantly larger” than fifty-five percent.⁸³

Numerical differences aside, the monopoly power must be durable.⁸⁴ Like with market dominance, there is no definitive proof indicating a business’s market share is enduring. But courts often look for evidence indicating barriers to entry—long-run costs new firms must pay that the defendant skirts—that would keep new competitors out of the relevant market.⁸⁵ Of relevance in the social media and social networking world are:

* Multi-Homing: Scott Morton and Dinielli (and others⁸⁶) claim that consumers are unlikely to switch from one social media platform to another and are less likely to use both Facebook and other platforms.⁸⁷ They believe that it is too burdensome for consumers to make the switch because all of their content and contacts are housed on Facebook already.⁸⁸ So as Facebook grows, more consumers become stuck on the platform, which in turn makes it difficult for a new competitor to attract users.

* Network Effects: Scott Morton and Dinielli (and others⁸⁹) likewise claim that social media platforms increase in value as they grow more popular.⁹⁰ The increase in users means an increase in content and contacts, which serves to attract even more users. The more people you know on Facebook, the more likely you are to join the site.

* Data Collection: Scott Morton and Dinielli (and others⁹¹) also claim that because Facebook and its peers can collect user data, they are able to improve their products and services based on consumer insights others do not have, and are able to attract ever-increasing revenues from advertisers, who value the data as a tool to target ads.⁹²

2. Step 2: Show Exclusionary Conduct

Even with this evidence in hand, plaintiffs are only halfway there. After the plaintiffs show monopoly power, they must show the defendant business acted in some exclusionary way to benefit its monopoly power.⁹³ In other words, plaintiffs must show conduct indicating “the willful acquisition or maintenance” of monopoly power that cannot be explained by “growth or development as a consequence of a superior product, business acumen, or historic accident.”⁹⁴ Distinguishing between illegal exclusionary conduct and legal competitive conduct requires analyzing harms to the competitive process itself.⁹⁵ Common harms include predatory pricing,⁹⁶ exclusive dealing,⁹⁷ refusals to deal,⁹⁸ tying,⁹⁹ and the like.

III. Facebook Isn’t a “Near-Monopoly” or Anywhere Close to It

Section 2 of the Sherman Act, you’ll recall, makes it unlawful to “monopolize, or attempt to monopolize, or combine or conspire with any other person or persons, to monopolize any part of the trade or commerce among the several States.”¹⁰⁰ Although section 2 creates three separate offenses,¹⁰¹ only monopolization is relevant here. Under that offense, plaintiffs must prove both (1) “possession of monopoly power in the relevant market”; and (2) “the willful acquisition or maintenance of that power as distinguished from growth or development as a consequence of a superior product, business acumen, or historic accident.”¹⁰²

Simply put, monopoly power is unlawful only when “accompanied by an element of anticompetitive conduct.”¹⁰³ According to Scott Morton and Dinielli, though, this is a difference that makes no difference: “Facebook,” they write, “has a monopoly in social media and/or social networks, whether considered in lay or legal or economic terms.”¹⁰⁴ In support of this conclusion, they first point to Facebook’s seventy-five percent market share in the “communications-focused” social media market.¹⁰⁵ Next, they give a laundry list of business decisions that are, from their vantage point, anticompetitive.¹⁰⁶

Neither argument can succeed. First, Scott Morton and Dinielli use the wrong market, which artificially inflates Facebook’s market share. And second, the businesses practices they cite as exclusionary are actually procompetitive.

A. Facebook’s Relevant Market

The first step is to define Facebook’s relevant market. At first glance, this task seems straightforward enough: Facebook is a social media platform, so the market must be social media platforms.

Yes and no. To begin with, Facebook is what economists call a “two-sided platform.”¹⁰⁷ This means Facebook “offers different products or services to two different groups who both depend on the platform to intermediate between them.”¹⁰⁸ In English, this means Facebook plays matchmaker between users and advertisers, and it competes for at least two different sets of customers. So, yes, Facebook competes in the social media platform market. But it also competes in the digital advertising market.

This follows from the rule that the market must include Facebook and all products or services that are “reasonably interchangeable” with it.¹⁰⁹ Other products or services are reasonably interchangeable when customers have the “ability and willingness” to turn to them following a price increase or “non-price change such as a reduction in [Facebook’s] product quality or service.”¹¹⁰ Market definition is thus “focuse[d] solely on demand substitution factors.”¹¹¹

So, what is Facebook and what are its products or services? In its legally required SEC filings, Facebook describes itself as:

* “[B]uild[ing] useful and engaging products that enable people to connect and share with friends and family through mobile devices, personal computers, virtual reality headsets, and in-home devices”;¹¹² and

* “[H]elp[ing] people discover and learn about what is going on in the world around them, enabl[ing] people to share their opinions, ideas, photos and videos, and other activities with audiences ranging from their closest family members and friends to the public at large, and stay connected everywhere by accessing [its] products.”¹¹³

The products Facebook alludes to include¹¹⁴:

* Facebook: a platform that “enables people to connect, share, discover, and communicate with each other on mobile devices and personal computers. There are a number of different ways to engage with people on Facebook, including News Feed, Stories, Marketplace, and Watch.”¹¹⁵

* Instagram: a separate platform that “is a place where people can express themselves through photos, videos, and private messaging, including through Instagram Feed and Stories, and explore their interests in businesses, creators and niche communities.”¹¹⁶

* Messenger: a “messaging application for people to connect with friends, family, groups, and businesses across platforms and devices.”¹¹⁷

* WhatsApp: a “secure messaging application that is used by people and businesses around the world to communicate in a private way.”¹¹⁸

* Oculus: the company’s “hardware, software, and developer ecosystem [that] allows people around the world to come together and connect with each other through [Facebook’s] Oculus virtual reality products.”¹¹⁹

From this bird’s-eye view, Facebook sounds a lot like a communications company. After all, its resume and product descriptions are chock full of industry buzzwords—“communicate,” “connect,” “messaging.” In fact, this language may have been on Wall Street’s mind when it recently revamped its Global Industry Classification System (the method used to match stocks with stock indexes) and bounced Facebook out of the “information technology” category into an expanded “communications services” category.¹²⁰ In its new home, Facebook joins the likes of AT&T, Verizon, and Comcast, as well as the also reclassified Google, Twitter, Snapchat, Netflix, and Disney.¹²¹

At first blush, this curious grouping may seem like Wall Street’s equivalent of the kitchen junk drawer. But with trillions of dollars on the line, the decision wasn’t made haphazardly; instead, it came after expert study of product and market evolution.¹²² As it turns out, though, one doesn’t need to be an expert to see what they see.

Think first of the broad similarities these companies share. In some form or another, each facilitates communication, distributes content, or does both. Now think of Facebook. As noted, the company’s products undeniably facilitate communication. But they also all distribute or help distribute content. Facebook lets users share and interact with user-made content (photos of the family vacation, videos of the new puppy, posts about almost everything), as well as content from third parties (a link to a GoFundMe page, a news clip from CNN’s Facebook page). Instagram likewise lets users do the same. For most people none of this is groundbreaking news—under the (unofficial) American Common Sense Classification System, Facebook is simply known as a social media platform.¹²³

But there is an important difference between the companies—their revenue streams. AT&T, Verizon, Comcast, Netflix, and Disney all have subscription-based services.¹²⁴ In other words, users must pay to hit “play” or “send.” Facebook, Twitter, Snapchat, and Google, on the other hand, are free for users to use. To be sure, advertising revenue is neither unique to these platforms nor shunned by their subscription-based peers. In fact, some offer multitiered subscription packages that offer lower prices on subscriptions but with more advertisements on their platform.¹²⁵

Free-to-use platforms rely so heavily on advertising, however, that they are said to have two sets of consumers: users and advertisers. As mentioned, because Facebook provides services to both, and thereby enables interactions between them, it is a two-sided market.¹²⁶ Two-sided markets usually have direct or indirect network effects.¹²⁷ Direct network effects occur when a product’s value increases with the number of people using it.¹²⁸ Indirect network effects, on the other hand, occur when the product’s value to one group increases the more another group uses it.¹²⁹

In plain terms that means Facebook operates in distinct, though overlapping, markets at the same time. First, it competes for users in the social media market. And second, it competes for advertisers in what’s known as the digital display ads market. Facebook’s legal filings recognize this too: “We compete with companies that sell advertising, as well as with companies that provide social, media, and communication products and services that are designed to engage users on mobile devices and online.”¹³⁰

Why does any of this matter? Well, if Facebook is a two-sided market with direct network effects, then the relevant market must be defined so as to include both Facebook’s social media market and its digital display market.¹³¹ But if Facebook has only indirect network effects, it may not be necessary to combine the markets.

Scott Morton and Dinielli assume the latter. They note early on that Facebook has “at least two types of customers,” and that advertisers “may have a location to display ads that is a substitute for social media.”¹³² But without much explanation, Scott Morton and Dinielli analyze Facebook as having only indirect network effects. This is a curious decision given statements they make throughout their report. In their section about entry barriers, for example, Scott Morton and Dinielli note that Facebook has strong “network effects,” explaining that the platform’s value increases the more users who use it.¹³³ They note also that Facebook’s value to advertisers increases with more users.¹³⁴

To be sure, Facebook may not have direct network effects. Although Facebook’s value to advertisers undoubtedly increases with the number of users who use the site (a bigger audience for them to reach), it is not entirely clear whether the same is true for users. As Scott Morton and Dinielli point out, Facebook’s value to users may even decrease as the number of ads increase.¹³⁵ Fair enough. But that really makes sense only if users must know the “how” or “why” behind Facebook’s business decisions in order to benefit from them. As a general matter, Facebook grows more valuable to users the more Facebook improves and innovates its products. Because Facebook’s revenue comes almost entirely from advertising revenue, ads are what fuel the company’s efforts to improve the user experience.¹³⁶ They also allow Facebook to remain free for users.¹³⁷ So although users may not know how advertisers benefit them, or why more ad revenue for Facebook can benefit them, they would feel the adverse effects if ads declined.

Because it is likely that Facebook has direct network effects, the relevant market must integrate the markets for social media platforms and digital display advertising. Here, there can be little doubt that Facebook is not a monopoly. Not only does Facebook compete against platforms like Twitter and Snapchat, it also competes against Google, Amazon, and a whole host of other platforms that place ads on their sites.¹³⁸

But even if Scott Morton and Dinielli are right that Facebook has only indirect network effects, the company is not a monopoly in either the social media or digital advertising market. Twitter, Snapchat, YouTube, and newcomer Parler—to name but a few—all compete with Facebook in the social media market. And, as just mentioned, Google, Amazon, and others all compete with Facebook in the digital advertising market.

Scott Morton and Dinielli try to circumvent this reality. Although they concede that Facebook operates in the social media market, they go on to define the relevant market so narrowly that their definition might as well be “Facebook and only Facebook.” In their telling, social media platforms “differentiate themselves from one another in various ways,” which means they actually occupy distinct markets.¹³⁹ One market includes “content-focused” platforms that “facilitate the distribution and consumption of content.”¹⁴⁰ Here, they lump YouTube and TikTok together because their content “can be enjoyed by users with a wide range of relationships to the person posting, including complete strangers.”¹⁴¹ That distinction doesn’t even make sense: like Facebook, TikTok, YouTube, Snapchat, and many others all allow public and private postings.

The other market includes “communication-focused social networks.”¹⁴² This, we’re told, is where Facebook belongs because it “primarily facilitate[s] communication (including the sharing of third-party content) among groups of friends who choose each other and enjoy content in large part because of those relationships.”¹⁴³ But even under this narrow definition, Facebook still faces competition from Snapchat and Twitter.¹⁴⁴

Here, any difference actually makes no difference. First, Scott Morton and Dinielli’s justification underscores that these platforms all operate in the same market. Why would the platforms need to “differentiate themselves from one another”—in one way or “in various ways”—if they didn’t compete against each other? Second, any validity the distinction between communication-focused and content-focused platforms may once have had no longer exists. As Scott Morton and Dinielli and the CMA point out, social media has blurred the lines between video sharing, messaging, blogging, and other services.¹⁴⁵ Their Roadmap implicitly acknowledges that Facebook is just as much about content as it is about communication. As discussed later, the Roadmap is full of criticisms about the content users see on Facebook’s platform.¹⁴⁶ Scott Morton and Dinielli go so far as to paint Facebook as parading increasingly harmful content before users to keep them engaged and ad revenue flowing.¹⁴⁷

In fact, the relevant market is probably even larger than just social media platforms. As Scott Morton and Dinielli unwittingly prove elsewhere in their paper, the relevant market should include all websites that compete for users’ attention.¹⁴⁸ After all, logic suggests that advertisers run ads on Facebook precisely because they expect Facebook users to see those ads. It therefore follows that if those users are on other websites, then those websites compete with Facebook (and its advertisers) for users’ attention.

So, it seems that Scott Morton and Dinielli are trying to have their cake and eat it too. As they would have it, Facebook is conveniently a communications-based platform when it comes to defining the relevant market. But when it comes to consumer harms, it is a content cesspool. To be sure, Facebook could be a communications platform that also contains harmful content. But that observation serves only to further weaken their arbitrary dividing line—if anything, it shows that Facebook is both communication and content focused.

Regardless of Scott Morton and Dinielli’s intent, their market definition ignores market realities. Facebook and its competitors have all evolved since their launch dates—a fact reflected in Wall Street’s decision to group these companies under the umbrella of “communication services.”¹⁴⁹ They also assume, without justification, that social media is static, that tomorrow’s platforms will look, feel, and act largely like today’s do. But it’s exactly that kind of thinking that led analysts, scholars, and the press to believe that Yahoo! Search was so dominant that it would never be beat.¹⁵⁰ Ditto for MySpace.¹⁵¹

In short, Facebook is a two-sided platform that must compete for users and for advertisers. Thus, at its broadest, Facebook’s relevant market is the “attention” market; at its narrowest, it’s social media. Whether it has direct or indirect network effects, and whether defined as a social media platform or a communications-based platform, Facebook still faces competition.

B. Market Share

With the contours of Facebook’s relevant market defined, we turn next to its market share. As you’ll recall, the Supreme Court has never found monopoly power when a company’s market share is below seventy-five percent.¹⁵² Fortuitously, Scott Morton and Dinielli’s analysis of Facebook’s market share happens to be exactly seventy-five percent.¹⁵³ But, like their definition of the relevant market, their analysis of Facebook’s market share is a game of analytical Twister.

First, a note about Scott Morton and Dinielli’s methodology. In traditional markets, market share is typically calculated by comparing a company’s sales (or the value of those sales) to all sales in the relevant market for a single year.¹⁵⁴ But because Facebook operates in a two-sided market and because it is free for users, another method is needed. So, Scott Morton and Dinielli and the CMA calculate market share based on the number of minutes a user spends on a particular platform divided by the number of minutes the user spends on social media in total.¹⁵⁵ They justify this metric on the grounds that companies “want to keep eyeballs on their platforms for as long as possible in order to sell more ads.”¹⁵⁶

This observation about a platform’s business model reinforces that Facebook operates in an integrated, multi-sided market. As a practical matter, the relevant market is broadly about attention—Facebook and other platforms all compete for users’ attention, which is what they sell to advertisers.¹⁵⁷ One way of keeping users’ attention is by making it as easy as possible for them to visit the platform—by, say, making it free to use. And although large financial outlays to attract and maintain customers are not unheard of in business, it is often unsustainable in the long-term. Thus, Facebook depends on advertisers for revenue to keep its platform attractive to users, and users keep the platform attractive to advertisers.

This feedback loop reveals that Facebook’s rivals are, at least, other online businesses that draw user attention through free products or services and that rely on digital advertising. Because user time is a scarce good—everyone alive gets the same twenty-four hours in a day—any time spent on another website is time not spent using Facebook’s products.

Second, given this reality, the market must include all social media platforms that operate like Facebook does. So, for example, Facebook’s share of user attention would be compared to that of Pinterest, Twitter, TikTok, Snapchat, YouTube, and other platforms. Scott Morton and Dinielli, however, define the market narrowly and exclude YouTube, even though it’s one of Facebook’s largest competitors for user attention and digital display advertisers.¹⁵⁸ With that competitor gone, Scott Morton and Dinielli peg Facebook’s market share—in the United Kingdom, remember—at seventy-five percent.¹⁵⁹ To arrive at this number, they added Facebook’s, Instagram’s, and WhatsApp’s market shares together.¹⁶⁰

Catch all that? First, Scott Morton and Dinielli choose a metric—eyeballs on ads—that directly undercuts their narrowly defined market. Unfazed by this bait and switch, they plow ahead with their formula—subtract YouTube, add a heaping of Facebook-owned products, and voila, seventy-five percent market share!

This makes little sense. If YouTube, which competes with Facebook for advertising dollars is out, then why is WhatsApp, which doesn’t run ads, in? And how is Instagram more communication based than content based? Even worse, Scott Morton and Dinielli aren’t even using market shares for the right year. The CMA’s own chart, which is reproduced below, shows that Facebook—even when all its products are combined—has seen a dramatic decline in market share since 2015.¹⁶¹ And looking at one year, the chart shows that Facebook’s combined market share hovers around forty percent.¹⁶² Regardless, the CMA’s own chart shows that Facebook’s market share is not durable; it’s down twenty percent since 2015.¹⁶³ In other words, even if Scott Morton and Dinielli are right that Facebook is dominant, that dominance is not durable.

Perhaps recognizing that this formula is too removed from common sense, Scott Morton and Dinielli try to buttress their finding. Their additional points include:

* Citing the CMA’s finding that the market gained only two new competitors—Snapchat and Instagram—over the last few years;¹⁶⁴ and

* Claiming that no competitor “has achieved significant market share.”¹⁶⁵

Even under their narrowly defined market, the first point is not sound. Microsoft’s LinkedIn has seen massive growth;¹⁶⁶ so, too, with the newly launched Parler platform.¹⁶⁷ And under a properly defined market, TikTok must be added to the list. The second point also misses the mark. Even if that were true, antitrust law doesn’t require competitors to achieve “significant” market shares to count as competitors.¹⁶⁸

Next, Scott Morton and Dinielli note that “[m]any people use more than one social network,” which is called “multi-homing.”¹⁶⁹ That’s certainly true. But they contend that this supports, rather than undermines, their argument that Facebook has a “near” monopoly on social media.¹⁷⁰ For proof, they cite these statistics:

* Ninety-seven percent of Instagram’s users visit Facebook but only sixty-six percent of Facebook’s users visit Instagram;¹⁷¹

* Ninety-five percent of Snapchat users visit Facebook, but only sixty-eight percent of Facebook’s visit Snapchat;¹⁷² and

* Seventy percent of TikTok users visit Facebook.¹⁷³

These statistics, we’re told, “confirm” that Facebook is “likely to have market power,” and that because TikTok’s users, who “skew young,” still use Facebook, users “do not view this platform as a substitute for Facebook.com.”¹⁷⁴

But these statistics, picked by Scott Morton and Dinielli themselves, reveal that Facebook does not have a near monopoly on social media. Instead, they show that a majority of Facebook’s users use other social media platforms, which seems to undercut their narrative that Facebook holds an undue sway over users’ attention. And although the multi-homing rates may not be perfectly reciprocal, they don’t have to be. Indeed, there are many reasons the proportion is not one to one—and none of them reflects anticompetitive or illegal behavior. For example:

* As Scott Morton and Dinielli note, TikTok’s users “skew young[er]” than Facebook’s users.¹⁷⁵ So if these users want to keep in touch with older relatives, they might turn to Facebook for that sole purpose.

* TikTok is also, as Scott Morton and Dinielli note, “the new kid on the block,” whereas Facebook opened to the general public in 2006.¹⁷⁶ Perhaps like Facebook in its early years, TikTok is embraced first by the young and later adopted by older users too.

* The same variables apply to Snapchat. It’s been around since 2011, and seventy-three percent of 18- to 24-year-olds use it, but only nine percent of 50- to 64-year-olds do.¹⁷⁷ By contrast, sixty-eight percent of the latter group use Facebook but only fifty-one percent of the younger group do.¹⁷⁸

Generational differences are worthy of elaboration. Facebook is popular among all age brackets, but it is most popular among those aged 18 to 29 and 30 to 49; it is least popular among those over 65 and under 18.¹⁷⁹ With these facts in mind, we can observe the following:

* Someone who is 29 today would have been 15 when Facebook launched in 2006. Back then, the only other major social media platform was MySpace, which was then thought to be so dominant that Facebook stood no chance.¹⁸⁰

* Today, someone who is 15 can choose between or among Facebook, Twitter, TikTok, YouTube, and Snapchat—to name just a few.

For added context, remember that Facebook launched during President George W. Bush’s second term; TikTok launched in the United States in 2018.¹⁸¹ Facebook allows for the sharing of political content; TikTok intentionally limits political content.¹⁸² TikTok had around 20 million monthly users in late 2018; it boasts more than 65 million today.¹⁸³ Based on these data points, we can speculate that perhaps TikTok is more popular than Facebook with young Americans because it’s new (and therefore intriguing) or because it’s not saturated with political content. And maybe TikTok is less popular among older Americans precisely because it’s new (and therefore relatively unknown or not appealing) and not saturated with politics.

Put simply, different generations of Americans have different tastes. It is thus unremarkable that rates of reciprocity are uneven—and, in fact, that suggests competition in the marketplace. If everyone used all social media platforms in equal proportion, then no platform would have as strong an incentive to innovate. Instead, TikTok has every incentive to eat into Facebook’s market share and to win over older Americans while cementing its popularity among the young. Conversely, Facebook must remain attractive to older Americans and try to compete with TikTok for younger ones.

To state the obvious, younger Americans are future consumers. Facebook must therefore win them over, or risk dissolving into irrelevancy. To be sure, younger Americans may switch to Facebook as they age. Perhaps they will come to desire more political content once they graduate high school. Or perhaps they’ll use Facebook solely for the purpose of reading political content while keeping TikTok as their main social media platform. Or maybe they’ll use Facebook, TikTok, Twitter, and other platforms for different purposes.

Whatever they may do, it is wrong to believe that social media platforms must resemble Facebook in order to be viewed as an alternative to Facebook. Scott Morton and Dinielli argue that although the market for social media includes platforms like YouTube, Facebook’s relevant market is actually narrower than that—it includes only “communication-focused” platforms that are, in all relevant respects, like Facebook. That narrow focus wrongly assumes, however, that American consumers view Facebook as the preferred “model” for social media. Seen this way, Facebook is of course going to dominate the market, just as Apple would be seen as dominating the smartphone market if we defined that market by features unique to iPhones.

Scott Morton and Dinielli also argue that a

significant reason that Facebook has market power is that a user cannot change platforms and expect to be able to stay in contact with her friends. Because social networks are not compatible, a user’s friends would have to change platforms with her for her to be able to continue to see their feeds.¹⁸⁴

That observation is true only if one defines “stay in contact with her friends” to mean “to stay in contact with her friends in exactly the same way as they stayed in contact on Facebook.” As a practical matter, it is true that if a user left Facebook, she would, well, leave Facebook. But why Facebook’s unique features should be available on Facebook’s competitors’ platforms is never explained. And, as it were, Americans don’t necessarily want social media to resemble Facebook. The mere existence of “multi-homing” suggests Americans don’t rely solely on Facebook for their social media needs, nor wish to confine themselves to Facebook even if it is the largest platform.

Scott Morton and Dinielli also give Facebook more credit than it deserves; indeed, they discount factors outside Facebook’s control. For many Americans, especially older Americans, Facebook is familiar. Given the ever-evolving nature of social media platforms, older demographics may use Facebook simply because they know it best. That’s not a result of any anticompetitive behavior; that’s just good luck from timing.

Facebook is also an old platform, at least relative to its competitors. Again, this timing probably matters: Generation Z came of age with many social media platforms, including TikTok, which is dominant among that demographic.¹⁸⁵ By contrast, Millennials, Generation Xers, and Boomers all started using social media when Facebook was not only new but really the only option.¹⁸⁶ Their sustained use of the platform is probably driven in part by comfort, convenience, and familiarity.

That some platforms focus heavily on certain forms of communication—like photos or videos—and Facebook offers more diversity does not mean Facebook represents what social media will look like in two, five, or ten years. Instead, it means that Facebook has a model that works for most of its current users. The key will be whether Facebook is able to attract new, younger users—and so far, it has been less successful.¹⁸⁷

So, if we were to follow Scott Morton and Dinielli’s assumptions, we’d calcify the social media market such that all options more or less looked like Facebook. That may be fine for many consumers, but the clear majority of Americans like the variety they have now. In any event, Facebook does not have dominant market share, let alone monopoly power.

C. Evidence of Monopoly Power

1. Direct Evidence

No matter the size of a company’s market share, plaintiffs must still prove that the defendant business has monopoly power.¹⁸⁸ One way of doing that is through direct evidence that shows a company charges prices far and above what a competitive market would support.¹⁸⁹ Scott Morton and Dinielli allege two pieces of direct evidence to show Facebook has monopoly power: (1) “social networks have strong direct network effects”; and (2) “Facebook has a near-monopoly share and enormous reach.”¹⁹⁰ Because Facebook is free for users, Scott Morton and Dinielli substitute qualitative factors like privacy protections and content type for price.¹⁹¹

Turning to their first argument, Scott Morton and Dinielli claim that “[a] very significant reason that Facebook has market power is that a user cannot change platforms and expect to be able to stay in contact with her friends.”¹⁹² Their logic boils down to this: social media platforms lack interoperability, which deters users from abandoning Facebook, which forces users “to put up with Facebook’s fake news, or exploitation of privacy, or any other increase in quality-adjusted price.”¹⁹³ Scott Morton and Dinielli further allege that these costs compound when the users looking to leave are members of a large group—if one member refuses to switch, he can block the group’s move.¹⁹⁴

That would be news to the nine percent of Americans who deleted their Facebook accounts in 2018,¹⁹⁵ the thirty-five percent who reported using the site less and less,¹⁹⁶ and the forty-nine percent of young Americans who forgo Facebook altogether.¹⁹⁷ Those who left the site entirely gave several reasons similar to Scott Morton and Dinielli’s suggestions, including the Cambridge Analytica scandal, the site’s echo-chamber nature, and the desire to be more productive—in other words, they were not forced to put up with anything.¹⁹⁸

Plus, Facebook already lets users export their data, including a list of their friends. Scott Morton and Dinielli highlight a user’s Facebook contacts as the most valuable asset on Facebook,¹⁹⁹ so this criticism seems misplaced. But to the extent that the list may not match another platform’s—for example, a friend that goes by one name on Facebook and another on Twitter—that is a problem not easily remedied by Facebook. Indeed, if Facebook were to allow users to export detailed identifying data on a user’s friends, Scott Morton and Dinielli would likely criticize the move as hurting user privacy.

Speaking of user privacy, Scott Morton and Dinielli’s quality adjusted price arguments are too subjective to use as proof of direct evidence. To be sure, US antitrust doctrine recognizes that nonprice factors can stand in for price increases.²⁰⁰ But Scott Morton and Dinielli do not make the case that fake news and privacy policies are such costs, let alone that they are above market level.

Take fake news first. Scott Morton and Dinielli cite it as a harm so widely understood as an objective harm that it needs no elaboration. But their discussion of fake news does not enjoy the widespread understanding that Scott Morton and Dinielli seem to take for granted: Republicans and Democrats don’t even agree on the definition of fake news.²⁰¹ Even if there were a common definition, many instances of fake news are in the eye of the beholder—what is fake to some is another’s alternative facts and is another’s reality. Practical problems aside, some users may also believe that, consistent with values of free speech, it is better for Facebook to allow more speech than less speech. Whatever the case may be, fake news is not objective enough to qualify as a cost forced down users’ throats.

The same is true of alleged privacy concerns. Polling reveals that Americans are worried about their digital privacy,²⁰² and reports show that Facebook has responded by giving users more control over their data privacy.²⁰³ If Facebook had monopoly power, it would not need to respond to these concerns at all. Plus, privacy operates on a sliding scale—some users want total privacy well others are content with sharing some or all data in return for free services.

2. Indirect Evidence

Plaintiffs may also show monopoly power through indirect evidence of entry barriers. Scott Morton and Dinielli provide many reasons why Facebook is protected from competition.²⁰⁴ But these arguments fare no better.

First, Scott Morton and Dinielli claim network effects are a “significant” barrier to entry for new firms.²⁰⁵ Network effects, you’ll recall, occur when a product’s value grows proportionally with the number of people who use it.²⁰⁶ Some scholars, including Scott Morton and Dinielli and the CMA, believe that when a platform’s network effects “tip”—meaning, it has so many users that it becomes the default and only real option for users—the market closes its doors to new entrants.²⁰⁷

At best, this is an unproven-but-possible problem. To date, there is no empirical evidence that social media platforms have “tipped.”²⁰⁸ But there is evidence that it does not currently plague the market. Just this year, for example, a start-up platform called Parler entered the market and immediately attracted millions of new users—many of them conservatives who feel Facebook, Twitter, and other platforms moderate too strictly.²⁰⁹ TikTok, meanwhile, entered the market in 2018 and is now the most popular app among young Americans.²¹⁰

What’s more, only fifty-one percent of young Americans use Facebook,²¹¹ but supermajorities of them use other platforms like Snapchat and TikTok.²¹² If the network effects posed an actual barrier to entry, we’d at least expect to see these young consumers use both Facebook and its competitors—indeed, Scott Morton and Dinielli claim that because Americans more broadly use other platforms and Facebook, that’s proof of Facebook’s monopoly.

Second, Scott Morton and Dinielli claim that Facebook makes it difficult to multi-home because it does not offer full API access to all its competitors.²¹³ This argument is false. Facebook makes its API accessible to competitors and is used by direct competitors such as Tinder, Viber, TikTok, Zoom, and Pinterest.²¹⁴

This argument also falls apart under common sense. Americans multi-home all day: many check Twitter, Instagram, TikTok, Pinterest, and Reddit regularly. Even Scott Morton and Dinielli acknowledge this. As they report, many Americans use multiple digital platforms.²¹⁵

Their complaint, however, is that multi-homing is not “necessarily easy.”²¹⁶ In support of this, Scott Morton and Dinielli again cite the alleged lack of inoperability; this time claiming that Facebook does not allow users of products other than its own to simultaneously post content on another platform and on Facebook’s platforms.²¹⁷ Of course, as they recognize, this is in part “intrinsic to the fact that most platforms are, indeed, run independently from one another.”²¹⁸ Part of that independence includes content-moderation policies. Facebook may think that such simultaneous posting could cause more headaches than it is worth. Two points jump to mind. First, Facebook already has billions of pieces of content to worry about. Adding even more, especially when the platform is already under scrutiny from the left and right, would likely require the platform to hire even more employees to review flagged content and handle disputes. Second, it could lead to even more politicized attacks. Suppose someone’s tweet posts on Twitter and Facebook, but Twitter removes it for violating its policy. If Facebook decides not to remove the content from its platform, it will surely invite even more criticism from certain corners. The same is true in reverse.

Facebook also works with third parties to help them use its APIs in ways that respect users’ expectations.²¹⁹ Despite Scott Morton and Dinielli’s claim to the contrary,²²⁰ thousands of publishers, websites, and competitor platforms have access to the site’s APIs.²²¹ But even if that were not the case, Scott Morton and Dinielli do not explain why Facebook must share its technology with its competitors. Facebook is not an essential facility, even if they believe that it is.

Third, Scott Morton and Dinielli claim Facebook’s data collection and use practices gives it a massive leg-up on its advertising competitors.²²² Data is neither finite nor incapable of being shared with one entity at any given time. In other words, Facebook may have data useful to its advertisers, but that data is not Facebook’s alone. Indeed, as Scott Morton and Dinielli reported in their first Roadmap, Google also has useful data for advertisers to capitalize on.²²³ So, too, with Amazon. Both Google and Amazon are Facebook’s largest competitors in the digital advertising space.²²⁴

Recent news also debunks Scott Morton and Dinielli’s claim. Some of Facebook’s largest advertisers have recently boycotted the platform in an effort to pressure the platform to change its moderation policies.²²⁵ If these advertisers were forced to use Facebook’s advertising tools, as Scott Morton and Dinielli believe,²²⁶ surely these boycotts would not happen—especially during coronavirus when retailers and manufacturers are already relying on online sales to compensate for declining in-person sales.²²⁷

Fourth, Scott Morton and Dinielli claim that Facebook’s reach makes it too valuable for advertisers to pass up.²²⁸ This is a particularly silly argument. The platform’s reach is valuable to advertisers—and that’s a benefit, not a barrier to entry. That advertisers can reach millions of potential consumers is a good thing, as discussed above.²²⁹ But, again, that does not mean Facebook has monopoly power.

IV. Facebook’s Conduct is Procompetitive & Benefits Consumers

A. Section 2 of the Sherman Act Requires Proof of Anticompetitive Conduct

In passing the Sherman Act, Congress sought to enshrine national values like competition and free enterprise.²³⁰ Those values usually go hand in hand,²³¹ but sometimes market realities make them seem like the odd couple.²³² On the one hand, the presence of a monopoly signals that the market lacks competition; on the other hand, striving for monopoly power “is an important element of the free-market system.”²³³ This apparent tension, however, merely reflects that competition and free enterprise are like a double helix. That is, “[t]he opportunity to charge monopoly prices—at least for a short period—is what attracts ‘business acumen’ in the first place,” and thus “it induces risk taking that produces innovation and economic growth.”²³⁴

Potentially earning monopoly profits is thus like potentially winning the lottery—remote though it may be, the mere possibility is temptation enough to try. But if that chance all but disappears, so too will the pool of eager risk takers. For similar reasons, as Judge Learned Hand explained, “[t]he successful competitor, having been urged to compete, must not be turned upon when he wins,” lest others be scared off from even trying.²³⁵ Accordingly, monopolies are unlawful only when “accompanied by an element of anticompetitive conduct”²³⁶ (translation: they can’t rig the lottery).

To meet this standard, plaintiffs must prove the monopolist “harm[ed] the competitive process and thereby harm[ed] consumers.”²³⁷ It’s not enough, in other words, to show that a monopolist hurt its competitors—that is to be expected because “[c]ompetition is a ruthless process.”²³⁸ It spares no business, not even big ones. Indeed, in just the last decade or so it has reduced to ruins market giants like Blockbuster, Borders, Circuit City, and Radio Shack.²³⁹ Once ever present, these companies have since closed up shop or filed for bankruptcy.²⁴⁰ And competition seems poised to shutter shopping mall staples like Victoria’s Secret, GameStop, GNC, and JCPenney next.²⁴¹ But the Sherman Act stomachs²⁴²—even cheers—this ruthless competition because “[a]ggressive, competitive conduct by any firm, even one with market power, is beneficial to consumers.”²⁴³

Harder to digest is “[a]ggressive, exclusionary conduct [that] is deleterious to consumers.”²⁴⁴ This is the sort of conduct that softens the market’s sharp edges and reduces pressure on a company to innovate.²⁴⁵ This behavior usually means the offending company is shielded from “competition on the merits.”²⁴⁶

But procompetitive and exclusionary conduct are often brewed in the same barrel: conduct that benefits consumers also tends to exclude competitors.²⁴⁷ So even with an eye toward consumer welfare, spotting the difference can be difficult.²⁴⁸ This is especially true in two-sided markets, where a platform’s conduct may benefit one group of consumers, while seemingly harming another.²⁴⁹ Even more than other areas of antitrust law, this inquiry eludes bright-line rules because, as Justice Scalia put it, “[b]ehavior that might otherwise not be of concern to the antitrust laws—or that might even be viewed as procompetitive—can take on exclusionary connotations when practiced by a monopolist.”²⁵⁰ For these reasons, a claim of anticompetitive conduct must be “judged against the realities” of the monopolist’s relevant market.²⁵¹

Judged against market realities, Scott Morton and Dinielli’s claim that Facebook’s aggressive competition amounts to unlawful exclusionary conduct falls apart. Even if Facebook were a monopoly, their arguments rest on inference built upon inference; not market realities or concrete facts. To be sure, they warn readers upfront that they “have not conducted [their] own factual investigation of Facebook,” and rely instead on the CMA’s preliminary report and “a small number of additional public sources, to map out what a monopolization case against Facebook in the US might look like.”²⁵² Still, Scott Morton and Dinielli are so sure of the evidence they marshal—limited as it is—that they conclude with absolute certainty that “[t]he facts that are public today do not provide a procompetitive justification for any” of Facebook’s business decisions.²⁵³

This conclusion is a curious one, not least because their own Roadmap admits that it draws speculative conclusions from incomplete facts. Start with the Roadmap’s “Anticompetitive Conduct” section, which opens with a hat tip to Justice Scalia²⁵⁴: “Conduct that might be relatively harmless in a different sort of market—the acquisition of a small, nascent competitor or complement, for example—might be dangerously anticompetitive in the social network market.”²⁵⁵ Although the “dangerously” qualifier seems over the top, the “might be” renders the statement possibly true. As it were, the Roadmap gets a lot of mileage out of this theoretically could-be-true method of analysis. Subsequent paragraphs are full of qualifying phrases like “may have,” “can be,” “in our view,” “what one would expect,” “given the uncertainty,” “would have done,” “[i]t is not clear,” “might,” “appears,” “may,” “suggests,” and so on.²⁵⁶

All the same, like antitrust law professors, Scott Morton and Dinielli are free to speculate about a hypothetical antitrust case against Facebook for being a hypothetical monopolist that engages in hypothetical anticompetitive conduct. The Roadmap is free to use qualifying language like this Article does in some places (see, e.g., next sentence). But as if growing more confident with each page, the Roadmap drops fewer and fewer clues that its content is speculative or that the evidence is shaky. And when it reaches a crescendo in confidence, the Roadmap even overrules the uncertainty its underlying source projects, reporting it instead as a dead-certain conclusion.²⁵⁷

Regardless, semantics do not change the conclusion that Facebook’s behavior, though aggressive, is procompetitive. Or—to play it safe—Facebook’s behavior is not anticompetitive. Either way, the market realities show that Facebook has indeed competed on the merits and has done so quite successfully. And in the process, it has benefitted consumers. To reach the opposite conclusion, as Scott Morton and Dinielli do, requires an analytical framework that has been tried by plaintiffs—and repudiated by courts.

B. Facebook’s Conduct

None of the claims Scott Morton and Dinielli raise about anticompetitive conduct are persuasive. First, they argue that Facebook “may” have bought up its competitors so that it did not have to compete on the merits.²⁵⁸ Second, they claim it “used its APIs [Application Programming Interfaces] as a tool to limit [users’] interaction with potential competitors and to weaken them.”²⁵⁹ Third, they argue that Facebook raised its rivals’ costs by hiding the true “cost” of its privacy policies.²⁶⁰ Fourth, that Facebook also hid those costs from consumers.²⁶¹ And finally, they argue that Facebook clearly intended to hurt its competitors.²⁶² These arguments are but a mere silhouette of antitrust law. They trace antitrust’s broad contours—and that’s about it.

1. Facebook’s Acquisitions

Scott Morton and Dinielli first attack Facebook’s acquisitions of Instagram, WhatsApp, and other complementary apps.²⁶³ According to them, Facebook paid “significant premiums over what one would expect based on the targets’ financial results,” which suggests Facebook bought them only to preserve its monopoly.²⁶⁴ In other words, Facebook wasn’t just some poor schmuck swindled into overpaying; it was a monopolist with an envious eye on Instagram’s “popularity with young people and its large user base.”²⁶⁵ And on WhatsApp’s “huge user base, largely in markets outside the US.”²⁶⁶

To begin with, that argument ignores the market realities at the time of acquisition and the market realities of the present. Start with Instagram. Launched in 2010 and bought in 2012, the photo-sharing app had no independent revenue streams, only a handful of employees, and around 30 million users.²⁶⁷ At 30 million, its user base was hardly “large,” especially not at a time when Facebook’s was nearing a billion.²⁶⁸ But Instagram must have had potential: it was valued at $500 million by investors and Facebook doubled the ante, paying $1 billion for it.²⁶⁹ Today, Instagram is a full-fledged social media platform that boasts 116 million users in the United States and a billion worldwide.²⁷⁰ And with the help of Facebook’s tools, it generated $20 billion in advertising revenue last year.²⁷¹

Facebook’s gamble obviously paid off. But because Facebook seemingly overpaid at the time, Scott Morton and Dinielli believe the “purchase was not motivated only by Instagram’s earning potential.”²⁷² Probably. After all, Facebook bought Instagram in April 2012—a month before Facebook was to go public.²⁷³ In the lead up to its initial public offering, Facebook surely sought to address weaknesses that could have dampened investor enthusiasm and to maintain enthusiasm about the company’s prospects by creating a superior and integrated social media platform. For starters, news stories about the acquisition mentioned Facebook’s need to break into the market for mobile apps.²⁷⁴ But why overpay? Well, it seems Facebook did not believe it was overpaying, a judgment later proved correct. That others missed what Facebook saw does not support Scott Morton and Dinielli’s conclusion that Facebook sought only to kneecap its potential competitor—it reflects Facebook’s business acumen. Plus, it is entirely possible that Facebook thought such a move would signal to investors that it was aware of its need to stay innovative and would do what was necessary to keep delivering superior products and services to users.

But even if Scott Morton and Dinielli are correct, Facebook’s acquisition of Instagram was still procompetitive. Under Facebook’s leadership, output increased dramatically: as noted, Instagram’s user base jumped from 30 million to one billion and its revenue rose from zero to $20 billion. Instagram’s quality improved too, upgrading from a photo-editing and photo-sharing app to a social media platform.²⁷⁵ And because increased output is generally recognized as a procompetitive effect,²⁷⁶ that alone disproves Scott Morton and Dinielli’s claim that the acquisition lacked any procompetitive effects.²⁷⁷

Even so, the Roadmap tries to skirt these market realities by claiming that “competition is beneficial to consumers even if the entrant eventually fails.”²⁷⁸ In other words, even though Instagram may have failed absent Facebook’s purchase—remember, it had no revenue streams—that still would have benefitted consumers more than what actually happened. Thus, Scott Morton and Dinielli appear to believe that, when faced with actual evidence of consumer benefits and imagined possibilities of even greater benefits, the latter should win the day. That’s hard to swallow. Regardless, Facebook’s acquisition of Instagram was “precisely the sort of competition that promotes the consumer interests that the Sherman Act aims to foster.”²⁷⁹

So too with WhatsApp. As Scott Morton and Dinielli describe it, WhatsApp is “principally an encoded messaging app” that is popular “largely in markets outside the US.”²⁸⁰ And like Instagram, it had “virtually no revenue stream” at the time of acquisition.²⁸¹ But that did not stop Google from offering $10 billion to buy WhatsApp, or Facebook from ultimately paying $19 billion for it.²⁸² Because Facebook seemingly overpaid for yet another revenue-raising dud, and has yet to successfully monetize it, Scott Morton and Dinielli again conclude that Facebook must have done so for anticompetitive reasons.²⁸³ Or, as is likely the case, Facebook saw in WhatsApp what it saw in Instagram: potential to improve iown products and services for users.

Consider the market realities. When Facebook bought WhatsApp in 2014, the company was still playing catch up in the mobile-messaging race, especially overseas.²⁸⁴ Facebook saw WhatsApp as a way to remain competitive in that race, and had been in talks with WhatsApp about a possible acquisition for at least two years.²⁸⁵ So although the purchase had the effect of keeping WhatsApp from falling into Google’s hands, it also kept Facebook from having to bow out. But killing two birds with one stone does not make for purely anticompetitive conduct. Indeed, if Facebook merely wanted to prevent competition, it failed big time. Not only was the mobile-messaging market highly competitive in 2014,²⁸⁶ it remains so today.²⁸⁷

Despite Facebook’s struggle to monetize WhatsApp, the acquisition still reflects the company’s sound business judgment. As just mentioned, it threw Facebook a life preserver while it struggled to stay afloat in the mobile-messaging market. But it also seems part of Facebook’s long-term strategy to move toward private messaging services both at home and abroad.²⁸⁸ Either way, the acquisition had a procompetitive rationale. And it’s had a procompetitive effect: like Instagram, its output has increased (to the tune of 2 billion users)²⁸⁹ and ditched its subscription fee to become entirely free.²⁹⁰

All in all, Facebook’s acquisitions are exactly the sort of procompetitive conduct that the Sherman Act encourages. Although Scott Morton and Dinielli try their best to cast Facebook’s acquisitions in anticompetitive garb, market realities and concrete consumer benefits make clear that Facebook’s conduct is lawful. Indeed, they underscore that Facebook is determined to avoid MySpace’s fate by addressing its market weaknesses before it’s too late and by responding to market trends early.

2. Facebook’s Application Programing Interfaces

Next, Scott Morton and Dinielli seem to argue that Facebook has a duty to deal with its rivals and has breached that duty. First, they argue that “the most realistic path to entry [for a new firm] may be through beginning as a complement [to Facebook], developing a user base and brand recognition, and if that is successful, later transitioning to an independent platform.”²⁹¹ Second, they conclude that “Facebook has used its [Application Programming Interfaces] as a tool to limit interaction with potential competitors and to weaken them, and to advantage itself.”²⁹² And third, they suggest that regulators consider requiring Facebook to “interoperate” with new rivals.²⁹³

Neither their analysis nor proposed remedy is persuasive. To begin with, section 2 of the Sherman Act is usually enforced through negative duties—that is, the duty not to act anticompetitively. But decades ago, in Aspen Skiing Co. v. Aspen Highlands Skiing Corp.,²⁹⁴ the Supreme Court held that in very narrow circumstances, a monopolist may have an affirmative duty to assist rivals.²⁹⁵ In that case, the Court affirmed a jury verdict against a ski operator who owned three out of four mountains in the area and ended a deal with the owner of the fourth mountain that let skiers access all four mountaintops.²⁹⁶ The Court held that jury’s verdict was reasonable because the defendant: (1) ended a profitable deal; (2) refused to even sell its tickets at market rate to the competitor; and (3) offered no legitimate reason for doing either.²⁹⁷ Although that case has never been overruled, the Court narrowed its application, characterizing it as being “at our near the outer boundary” of section 2’s scope.²⁹⁸

Even if the Court were to resurrect Aspen Skiing, applying its rationales to Facebook goes downhill quickly. As a general rule, the Sherman Act affords all companies—even monopolies—great discretion in choosing with whom to do business.²⁹⁹ That discretion flows from the Sherman Act, which protects the competitive process for the benefit of consumers, not for the fortunes of competitors.³⁰⁰ And it reflects the reality that a duty to deal is likely to become a right to share, which is inconsistent with competition.³⁰¹ It also risks forestalling, rather than promoting, competition by reducing incentives for competitors to develop their own inputs.³⁰²

Scott Morton and Dinielli have not overcome the strong presumption that Facebook is free to deal or not deal with its competitors as it sees fit. Indeed, Facebook’s platform may be useful for rivals—as they note, it gives them Facebook’s user base and social graph³⁰³—Facebook simply doesn’t have dominant control over finite resources like the defendant in Aspen Skiing. As trite as it may sound, it’s worth remembering that, unlike four mountains in one region, Facebook is but one platform on the internet. The internet spans the entire world and can host as many platforms as society can imagine. At bottom, Facebook has “no general duty to help its competitors”; it doesn’t even have to “extend a helping hand to new entrants.”³⁰⁴

Even under the duty-to-deal offshoot “essential facilities” doctrine developed by a few lower courts, Facebook’s decisions about API access would still be found lawful. Under that doctrine, plaintiffs must prove four things: (1) the monopolist controls access to an “essential” facility; (2) competitors cannot “practically or reasonably” duplicate the facility; (3) the monopolist denied a competitor access to its facility; and (4) the monopolist can “feasibly” share access to the facility with its competitors.³⁰⁵ But even when a plaintiff passes the doctrine’s test, evidence of “legitimate business reasons” for restricting or refusing access will defeat the claim.³⁰⁶

Take Scott Morton and Dinielli’s Vine example. Facebook undisputedly controls access to the platform that bears its name. But, as even the Roadmap admits, Facebook competes with Twitter, which suggest the two products are reasonably interchangeable to begin with. And because Twitter has its own APIs, and because Twitter bought Vine, both Twitter and Vine had access to Twitter’s platform—in other words, it had access to a comparable facility.

But even if that weren’t the case, Facebook has legitimate reasons for deciding which apps get access and which do not. According to Scott Morton and Dinielli, Facebook’s decisions depend “largely on whether the app adds value to Facebook, which it must do to become a partner initially, or whether it poses a competitive threat, which is what causes the change in policy.”³⁰⁷ If an app doesn’t add value to Facebook, why would Facebook partner with it? Isn’t that cost-benefit analysis par for the course with all (successful) businesses? And even if it were true that competition concerns factored into Facebook’s decisions, Aspen Skiing makes clear that such a consideration may be a problem only when it’s the sole reason.³⁰⁸

With all that said, this is largely an academic debate. In reality, Facebook’s APIs are open to thousands upon thousands of third-party businesses.³⁰⁹ Facebook even grants access to its largest competitors, including YouTube,³¹⁰ TikTok,³¹¹ Pinterest,³¹² Tinder,³¹³ Zoom,³¹⁴ and Spotify.³¹⁵

Scott Morton and Dinielli may do things differently, but that does not make Facebook anticompetitive. And it certainly does not justify impeding Facebook’s right to do business with whom it chooses. In sum, Facebook’s decisions about API access aren’t anticompetitive.

3. Privacy, Data Practices, & Imagined Harms

Finally, Scott Morton and Dinielli argue that Facebook’s privacy policies and data-collection practices amount to exclusionary conduct because they hid what Facebook was doing behind the scenes.³¹⁶ Once again, neither argument is persuasive. To be sure, Facebook has gotten into hot water for some of its practices—the FTC fined it $5 billion in 2019 for misleading consumers about their privacy controls.³¹⁷ But the company’s conduct—at least as it relates to privacy and data collection—is not relevant to antitrust. To the contrary, that the FTC levied the largest fine in its history against Facebook and Facebook agreed to unprecedented reforms³¹⁸ underscores that these are trade practices governed by the FTC’s section 5 powers, not its antitrust division.

But even if Scott Morton and Dinielli are right that such conduct falls within antitrust’s ambit, Facebook still has not violated the law. Remember, the Sherman Act requires harm to the competitive process that in turn harms consumers. As this Section will make clear, the Roadmap is full of alleged harms—just not real ones.

Start with Facebook’s worst scandal: Cambridge Analytica. According to Scott Morton and Dinielli, that debacle reveals that Facebook “can gather, analyze, and sell information” about users.³¹⁹ All that data makes for better ads, and better ads means more advertisers—so, as they explain, Facebook gains “but consumers may not.”³²⁰ How might consumers be harmed? First, Scott Morton and Dinielli suggest that Facebook “accentuates or provokes emotions such as hate or fear.”³²¹ Second and relatedly, it promotes “outrageous and arousing” content so that consumers become “addicted” to the platform.³²² Third, Facebook can “exploit” all “these emotions to enable advertisers to sell products they otherwise would not, e.g., makeup to depressed teenagers.”³²³

In other words, Facebook’s targeting advertising is so good it hurts. But these “harms” are unrelated to antitrust at best, and entirely subjective and imagined at worst. After all, the behavior described by Scott Morton and Dinielli is not unique to Facebook. Indeed, politicians, producers, platforms—anyone with a basic understanding of persuasion—promote messages that they think will resonate with their target audiences every day. Nor are Facebook’s tactics unique: for centuries, companies have appealed to consumers through ads tied to emotions, values, or principles. A common example: “American Made” decals that signal to American consumers that by buying a certain product they are supporting their country’s workers and economy. Another example: food commercials that distinguish their “fresh-made” product from those of their unfresh competitors. So pervasive is this technique that even Scott Morton and Dinielli use it: consider their example of selling “makeup to depressed teenagers.”³²⁴ That example is meant to stir up a reader’s emotions—how dare a giant company like Facebook prey on a vulnerable teenager’s insecurities to make an easy buck on something as inconsequential as makeup!

Everyone may use persuasive messaging, sure, but what about messaging that appeals to “hate” or “fear”? Even assuming that consumers respond the same way to the same content, neither emotion is itself a consumer harm. And, in some cases, these emotions may actually benefit consumers. For example, if an ad scares the public into following public-health protocols during a global pandemic—by, say, displaying charts of rapidly rising infection rates and fatalities—many would see that as a net benefit.

To be fair, Scott Morton and Dinielli likely have in mind examples that are far more unseemly. But therein lies the problem: messages are inherently subjective. And subjectivity and antitrust do not pair well. Imagine the FTC, DOJ, and courts trying to define “outrageous” or “arousing” content objectively enough that they can spot when it crosses the line into consumer harm. Even if the government adopted a standard like the one hinted at by Scott Morton and Dinielli—content that makes consumers buy things they otherwise would not—it would still be at a loss. If a commercial of a juicy hamburger arouses a viewer’s appetite for fast food, even if such food is unhealthy and even if that person has perfectly good food in the fridge, is that commercial harmful?

That gets to another point. Let’s say commercials aimed at the general population are acceptable even if they’re arousing or outrageous, but the same ads cannot be targeted specifically at those for whom they’d resonate most. Most consumers like targeted advertising, so should that preference be overruled because, in Scott Morton and Dinielli’s opinion, they might buy stuff they don’t need? Should retailers that sell golf equipment stop sending catalogues to golfers?

Underlying all this is a fact that makes many, including Scott Morton and Dinielli, uncomfortable. That content might appeal so successfully to some users reveals that those users are predisposed to like that content to begin with. In other words, Facebook has an incentive to promote relevant content to its users. And because content can be relevant only in the eyes of the beholder, the content is a reflection of the best—and worst—of society. Take political content. Some Democrats—and advertisers—want companies like Facebook and Twitter to remove or otherwise address President Trump’s social media posts. Republicans, by contrast, want them to leave the content alone. Americans are also divided—some find the president’s posts objectionable; others do not. Who’s right and why? To be clear, platforms can decide for themselves when content violates their standards—when it is objectionable. Even better, users can decide for themselves, either by leaving a platform whose moderation policies they dislike or by advocating for change.

What cannot happen—not under the First Amendment³²⁵—is for the government to use its antitrust powers to decide which content is “addictive,” “outrageous,” or “arousing.” Suppose the First Amendment was out of the picture and the government focused only on addiction (an effect). How might it measure that harm? For example, the effects of nicotine addiction can be measured objectively. But how, exactly, would Facebook addiction make itself known? Take Sean and his mother, Margaret—fictional Americans who use Facebook. Say they both spend an average of three hours per day on Facebook. If Sean spends anywhere from two to three of those hours communicating with his friends, is he addicted and if so, is he addicted to Facebook or to talking to his friends? And if Margaret spends the same amount of time reading news articles and debating her colleagues about current events, what then? If one seems more harmful than the other, why? What if both report greater happiness after doing these activities because, presumably, they enjoy them?

Scott Morton and Dinielli’s complaint is that Facebook is uniquely problematic because its data may let advertisers create ads that are uniquely persuasive.³²⁶ Although advertising influences consumer decisions—if it didn’t, no one would bother advertising—it’s a stretch to claim that, somehow, in some way, Facebook is so persuasive and the harm is so severe that it can be measured and used for antitrust enforcement. That’s simply not the case. And, indeed, taken to their logical conclusion, Scott Morton and Dinielli’s arguments all rest on the government making subjective decisions about what’s good and what’s bad.

To be clear, this is not to say that the government cannot regulate Facebook’s conduct like the FTC did after Cambridge Analytica. But it is to say that the government cannot use its antitrust powers for perceived harms it cannot measure objectively. In any event, Scott Morton and Dinielli’s harms are purely hypothetical.

Conclusion

Clocking in at thirty-six pages, the Roadmap fails to make any credible antitrust case against Facebook. Indeed, even with antitrust enforcement as its destination, the Roadmap goes off course time and again. To be sure, it makes stops and speaks the language of antitrust—“market power,” “monopoly,” “anticompetitive,” “consumer harm,” and the like. But it does not give those concepts their due. To the contrary, the Roadmap takes irrelevant detours into subjective harms that clearly bother Scott Morton and Dinielli, but that are not illegal or harmful under antitrust doctrine.

Rather than making the compelling antitrust argument, the Roadmap instead advances policy arguments while advocating for antitrust reform. Under the cloak of the consumer welfare standard, the Roadmap tries to turn back the clock to when big meant bad. In the process, it reminds all readers that the revolution in antitrust that began in the early 1970s must once again be defended.

Facebook may be big—indeed, it may be very big—but it is no monopoly. Its products are among the most valued in the world precisely because they add so much value—for users, advertisers, and publishers. Scott Morton and Dinielli’s Roadmap recasts many of Facebook’s benefits to consumers as harms, and thus cleverly advances a big-is-bad, old-is-new-again approach to antitrust. And though that thinking may seem fresh today, the harm it will inflict on consumers will be a costly history lesson. Even so, if the country is convinced that this time will be different, that big really is bad, then perhaps change is inevitable. Any changes along the lines of Scott Morton and Dinielli’s arguments, though, should be seen for what they are: not as an extension of the consumer welfare standard—but as a repudiation of it.

Category: Print Issues

Debunking the "Big is Bad" Bogeyman: How Facebook Benefits Consumers

Share this article