Content moderation remains in the news following President Trump’s accusation that Google manipulated its searches to harm conservatives. Yesterday Congress held two hearings on content moderation, one mostly about foreign influence and the other mostly about political bias. The Justice Department also announced Attorney General Sessions will meet soon with state attorneys general “to discuss a growing concern that these companies may be hurting competition and intentionally stifling the free exchange of ideas on their platforms.” 

None of this is welcome news. The First Amendment sharply limits government power over speech. It does not limit private governance of speech. The Cato Institute is free to select speakers and topics for our “platform.” The tech companies have that right also even if they are politically biased. Government officials should also support a culture of free speech. Government officials bullying private companies contravenes a culture of free speech. Needless to say, having the Justice Department investigate those companies looks a lot like a threat to the companies’ freedom. 

So much for law and theory. Here I want to offer some Madisonian thoughts on these issues. No one can doubt James Madison’s liberalism. But he wanted limited government in fact as well as in theory. Madison thought about politics to realize liberal ideals. We should too. 

Let’s begin with the question of bias. The evidence for bias against conservatives is anecdotal and episodic. The tech companies deny any political bias, and their incentives raise doubts about partisan censorship. Why take the chance you might drive away millions of customers and invite the wrath of Congress and the executive branch on your business? Are the leaders of these companies really such political fanatics that they would run such risks? 

Yet these questions miss an important point. The problem of content moderation bias is not really a question of truth or falsity. It is rather a difficult political problem with roots in both passion and reason. 

Now, as in the past, politicians have powerful reasons to foster fear and anger among voters. People who are afraid and angry are more likely to vote for a party or a person who promises to remedy an injustice or protect the innocent. And fear and anger are always about someone threatening vital values. For a Republican president, a perfect “someone” might be tech companies who seem to be filled with Progressives and in control of the most important public forums in the nation. 

But the content moderation puzzle is not just about the passions. The fears of the right (and to a lesser degree, the left) are reasonable. To see this, consider the following alternative world. Imagine the staff of the Heritage Foundation has gained potential control over much of the online news people see and what they might say to others about politics. Imagine also that after a while Progressives start to complain that the Heritage folks are removing their content or manipulating new feeds. The leaders of Heritage deny the charges. Would you believe them? 

Logically it is true that this “appearance of bias” is not the same as bias, and bias may be a vice but cannot be a crime for private managers. But politically that may not matter much, and politics may yet determine the fate of free speech in the online era. 

Companies like Google have to somehow foster legitimacy for their moderation of content, moderation that cannot be avoided if they are to maximize shareholder value. They have to convince most people that they have a right to govern their platforms even when their decisions seem wrong. 

Perhaps recognizing that some have reasonable as well as unreasonable doubts about their legitimacy would be a positive step forward. And people who harbor those reasonable doubts should keep in mind the malign incentives of politicians who benefit from fostering fear and anger against big companies. 

If the tech companies fail to gain legitimacy, we all will have a problem worse than bias. Politicians might act, theory and law notwithstanding. The First Amendment might well stop them. But we all would be better off with numerous, legitimate private governors of speech on the internet. Google’s problem is ours.