Attacking “Big Tech” has become commonplace on the political right. Republican conference chair Elise Stefanik (R-NY) has grimly warned that tech platforms are taking the United States in an “authoritarian direction.” Media Research Center president Brent Bozell says that major platforms have “declared war” on the “American ideal of freedom.” Former Trump advisor Peter Navarro accused one tech company of being “in bed with the Chinese Communist Party;” Donald Trump, Jr., has charged another firm with violating the First Amendment (which it actually can’t do, since it’s not the government). Dozens of organizations have accused Big Tech firms of discriminating against conservatives and even pledged not to take money from them.
These appeals work because there is a widespread perception of tech companies’ anti-conservative bias. But just how bad is that bias?
A scanning of the landscape makes it clear that it’s not negligible. In March of 2021 Amazon stopped selling Ryan T. Anderson’s book, When Harry Became Sally, a scholarly and measured work, because some transgender groups found its arguments offensive—while continuing to sell Mein Kampf. Facebook hosts pages for antifa but has banned the Proud Boys (which, it is true, has been tied to more violence). Former President Donald Trump has been banned from Facebook and Twitter for promoting violence with his repeated lies about the 2020 presidential election; but the world is full of similarly truth-challenged people sewing discord like Venezuelan dictator Nicolás Maduro who retain their accounts.
The reasons for these distinctions are lost on many observers.
Yet even if a Big Tech anti-conservative bias is apparent, it’s not necessarily because of a purposeful suppression of viewpoints. Even neutral mechanisms administered with good intentions can have disparate impacts. Big Tech companies are almost all located in left-leaning areas like northern California; New York City; Austin, Texas; and northern Virginia, from which they draw their personnel. Big Tech’s desk jobs are populated with young, college-educated elites, among the most liberal demographics in the nation. When it comes to subjective matters like whether content is “offensive,” these employees, like everyone else, will tend to have more sympathy for information that affirms their pre-existing views, though many platforms have taken steps to limit these biases.
More important, tech platforms’ user bases skew younger and, therefore, more liberal. Most users of TikTok are under thirty, an age group that voted for Joe Biden over Donald Trump in 2020 by a twenty-five-point margin. In these circumstances, perfectly even-handed moderation would still likely treat the Left better than the Right if it relies on any amount of user reporting.
Cures Worse than the Disease
The reality and recognition of bias do not necessarily mean that public policy can solve the problem. Take the case of Anderson’s book on the transgender issue, for example: Amazon’s decision not to carry it doesn’t significantly restrict the author’s ability to share its message. In fact, the book is, like others, on balance much easier to get than it would be without Big Tech: Anyone with an internet connection can download it instantly or have a printed copy delivered the next day. Using government power to force any particular retailer to sell a book with whose message it disagrees is patently offensive.
Next, consider the reaction of Senator Elizabeth Warren (D-MA) to the decision by the Facebook Oversight Board to uphold the platform’s ban of Trump. She said she was “glad” that Trump had been banned from the platform, since he was “truly a danger to democracy.” She followed that, however, with a claim that Facebook itself was “too powerful.” That is, someone other than the privately owned Facebook or its independent Oversight Board should have banned the speech of a person who won more than seventy million votes in the most recent election. If a private company should not have that power, it follows that the government, now run by Warren’s co-partisans, should.
Those on the right have said things that raise at least as many questions. Senator Ted Cruz (R-TX), in a Wall Street Journal op-ed, declared that he would say, “No, thank you,” and let corporate-friendly programs expire when they involve corporations that take “woke” political positions with which he disagrees. The idea of using subsidies and regulations to reward or punish private business on the basis of political views, which is what Cruz implies doing, is the definition of crony capitalism and a severe threat to democratic norms. In fact, most actions that would force specific corporations, in technology or any other sector, to treat conservatives better are likely to be discriminatory and deeply anti-democratic.
The questions Big Tech raises with regard to speech are real enough. And certainly it’s worth taking seriously Francis Fukuyama’s observation that the platforms “have neither the capacity nor the legitimacy to act as arbiters of democratic political discourse.” While Fukuyama and others have suggested some tech-related proposals for how content can be filtered through fewer hands, our politicians have yet to think seriously about how to address the problem.
At a loss for how to grapple with what is essentially a cultural and political problem in a way that comports with democratic norms, some policymakers think there is an economic solution. The control that Big Tech firms exert over speech demonstrates that they are “too powerful,” and therefore must be cut down to size. This fuzzy thinking inevitably leads policymakers to turn to indirect tools to solve the speech problem—specifically, antitrust enforcement.
Yet in turning in the antitrust direction, policymakers risk unwelcome economy-wide consequences. They would be wielding far too blunt an instrument and challenging free-market principles that many conservatives have long recognized as important to both economic vibrancy and individual freedom. Major antitrust regulatory legislation raises fundamental questions in and of itself, because Big Tech is not a single entity—and because enforcement would require reconsidering nearly fifty years of antitrust law in a way that would reshape the economy.
It is difficult to figure out any common regulatory rules that would plausibly fit all or most Big Tech firms. First, not all Big Tech firms are necessarily big: Twitter isn’t a Fortune 500 company. The Wikimedia Foundation is a nonprofit with a smallish workforce. Nor do these companies all make markedly heavy use of technology. There are significant examples from corporate America—Walmart, Nucor, McDonald’s—of firms that have made innovative use of proprietary technology, whereas Facebook simply refined social media principles that were more than a decade old.
What does, say, Amazon, a massive retailer and cloud services company with no significant social media operations, have in common with Facebook, which is close to social media “pure play?” True, both have pervasive roles in American life. (But so does Coca Cola.) Both collect large amounts of data and allow individuals and small businesses to do things that were once possible only for big business. Yet all modern firms, in any industry and of any size, rely on the internet to operate and all of any size collect some customer data. A total computer system failure would paralyze ExxonMobil just as much as it would Google, and a prohibition on collecting consumer data would be as problematic for Kroger as for Amazon.
That is why significant action to limit the power of Big Tech companies would require antitrust enforcement embodying a fundamental—indeed, a reactionary—understanding of antitrust law. For some fifty years, policymakers from both parties, heavily influenced by scholars like Robert Bork and Alfred Kahn, have operated under a “consumer welfare” standard, which emphasizes the impact of market behavior on consumers and on competition itself rather than on individual competitors. Since major online firms like Google and Facebook provide consumers with no-cost services, which these consumers value highly, they provide an inarguable gain in consumer welfare.
Not So Modest
There is, of course, intellectual support for alternative communitarian approaches that are inherently skeptical of size. There are data to support the ideas that smaller businesses simply do some things better than larger ones, that local ownership is inherently better than international or even out-of-town ownership, that very large companies have too much power in labor markets, and that antitrust law ought to protect particular competitors because a larger diversity of firms is inherently better. Also, some argue that letting any private person accumulate more than a certain amount of wealth is wrong or that no private entity should be allowed to consolidate market power. These matters are open to democratic debate, and there is nothing inherently anti-democratic about the idea of antitrust enforcement per se.
That said, they raise huge problems of definition and classification. If Amazon is “too dominant” in retail sectors, with too much capacity for predatory pricing, is Walmart, which is bigger, an even greater problem? If Google ($180 billion in revenue) poses insuperable competitive obstacles to number two player Microsoft ($140 billion) in the market of internet search engines, how big would Microsoft have to be to compete effectively? How many competitors should exist? What economies of scale do we need?
These questions are hard enough to answer under a consumer welfare standard; they would be even harder to answer under altogether new standards. People who favor antitrust action against Big Tech are asking for a centrally supervised restructuring of nearly the entire American economy, not of Big Tech alone.
Reforming Section 230 of the Communications Decency Act of 1996 is another major regulatory overhaul that proponents assume would neatly target Big Tech without fallout. It currently stipulates that “No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.” Some contend that this standard—based on precedents that, among other things, protect booksellers from being sued because of books they sell—grants social media companies a special privilege. Section 230 was a major factor in the expansion of the internet in the United States: It permits platforms to allow users to publish—and the platforms to take down—user-generated material without undue hazard. Significant change to the law would upend the business model of nearly every major Internet player. By violating the reliance interest that has built a huge industry, getting rid of Section 230 would be a big deal, not a modest step.
Other “modest” proposals have equally evident externalities. A proposed ban on acquisitions by already large companies, for example, would dry up a major source of capital for promising start-ups. The implications extend far beyond the tech sector. And these are only the steps that stop well short of direct action against particular companies, let alone Big Tech writ large. Nearly all the important Big Tech players have national or international reach: Even the idea of letting states experiment is a potential regulatory nightmare.
Frustration among politicians over bias in policing content contributes to a knee-jerk reaction to want to break up Big Tech. Yet there is every reason to think the same sort of content discrimination would occur among a larger number of smaller companies. Policymakers need to be careful: We don’t want to use a rocket-propelled grenade to catch fish. Cases of discrimination against conservatives have certainly occurred. It’s not implausible to think they could be the type of market failure that tends to self-correct. In any case, government actions against particular private-sector acts of speech discrimination would themselves be highly discriminatory and would give government the kind of power that should worry people across the political spectrum.
If Big Tech is in need of policy solutions, the best ones, at this point, would be tailored federal regulatory measures that are narrow in the matters they affect but broad in the companies to which they apply. This approach currently applies in areas from essential health insurance benefits to mobile-home building codes to most types of securities sales. One potential area for action would be consumer data protection. Early action in this area could ameliorate the current patchwork of federal, state, and international (mostly EU) laws. Such new laws would affect Big Tech, of course, but could also affect areas like airline frequent flyer programs and local restaurants’ customer databases. Good or bad, an imperfect common standard is sometimes preferable to a patchwork of local ones even if some of those local standards are closer to an ideal.
True, there are precedents in American history for the kinds of major regulatory overhauls being considered for Big Tech. They are consistent with democratic norms. They might even please some of today’s conservatives. But they are hard to square with the American Right’s longstanding commitments to free markets and limited government.
Eli Lehrer is president of the R Street Institute.
American Purpose newsletters
Sign up to get our essays and updates—you pick which ones—right in your inbox.Subscribe