Only About 3.5 Percent of Americans Care About Democracy

A recent study reveals that vanishingly few voters will defect from a candidate who acts undemocratically.

An illustration of the Statue of Liberty made up of many dots
Getty / Katie Martin / The Atlantic

Imagine a candidate you like. This politician has everything: the right positions on taxes, abortion, foreign policy, immigration; sound judgment; enough personal probity to be trusted with your wallet, house keys, or email password. Now imagine that that candidate does or says something antidemocratic. For no particular reason, she shuts down polling stations. Or at a rally, she tells supporters that a particular journalist—standing over there, in the Men’s Wearhouse sport coat—is asking too many questions and might deserve to get rabbit-punched on the way to his car. Care to change your vote?

This purely theoretical scenario, which of course bears no relationship to anything that has happened or is happening in American politics, is the subject of an article in the American Political Science Review by Matthew H. Graham and Milan W. Svolik of Yale University. How much do voters really care about democracy? Nearly all Americans say democracy matters. But how many will actually punish their preferred candidate and withhold a vote when that candidate does something undemocratic?

Graham and Svolik’s answer: About 3.5 percent of voters will defect from a candidate whom they otherwise support, but who does something destructive of democratic norms. Those 3.5 percent come from the right and the left in equal proportions, but they tend to be moderates. (Self-described “independents”—those mysterious, yeti-like creatures who profess to have no political preference at all—vote slightly more like extremists.) If you value democracy, hug a moderate.

“If you just ask people whether they like democracy, there’s a social norm that says they have to answer yes,” Svolik told me. They have been conditioned since grade school to say “democracy is good, 10 out of 10—and we should also stop global warming and save the whales and whatever.” He and Graham surveyed 1,691 people and posed instead a version of the hypothetical question I asked above: You say you like democracy, but will you sacrifice other things you like on its behalf, by withholding your vote for a democracy-bashing candidate? “Some will, but the punishment is small,” Svolik said: those willing to vote for the opposing candidate often do so only if he is similar to the candidate they intended to support in the first place. That means partisanship encourages more antidemocratic action: Stronger partisans will let the thuggishness slide, if they would have to sacrifice more than a small portion of their positions. The greater the number of strong partisan voters and politicians, the smaller the punishment for violating democratic norms, and the more likely the norm-breaker is to get elected.

Then Graham and Svolik checked their survey data against an actual case: the Montana congressional election that pitted the Republican tech mogul Greg Gianforte against Rob Quist, a Democrat best known for playing the banjo and other stringed instruments. On the day before the election, Gianforte became irritated with a line of questioning by Ben Jacobs, a Guardian journalist, and threw Jacobs to the ground. Gianforte, who won the election, later pleaded guilty to assault. He is currently in Congress, and is likely to be elected the next governor of Montana.

On Election Day, voters at the polls knew about Gianforte’s violence—but those who mailed in their ballots early had not. And in this real-world scenario, given to us by the gods of political science, Graham and Svolik’s prediction came true: About 3.6 percent of Gianforte’s votes on Election Day vanished relative to the votes counted before voters knew he body-slammed Jacobs. Body-slamming a journalist makes a difference, but not much of one, and especially not to extreme partisans.

Three and a half percent does not sound like a very healthy number. It means that nearly everyone you know who says democracy is sacrosanct is basically lying—which probably includes you and me. In the privacy of the voting booth, we reveal our hypocrisy. The terms of the hypocrisy help explain how authoritarian rule begins, Svolik told me. Viktor Orbán in Hungary, Recep Tayyip Erdoğan in Turkey—these figures came to power in elections. “Why do people elect them in the first place” when they are showing signs of authoritarianism? Svolik asked. “The key to understanding that is to understand when people are actually willing to vote in favor of democracy—and against something they like.”

But there is good news. First, Svolik said, the hypocrisy is evenly distributed. Democrats seem to be just as willing to break democracy as Republicans. On one measure, Svolik said, Republicans do slightly better: They are more willing to respect the First Amendment rights of their opponents. “Republicans actually punish those Republicans who would want to bar far-left protests,” Svolik said. Democrats also punish Democrats for barring far-right protests, but not as much.

Second, Svolik said, we have to bear in mind that democracy is not the only thing voters should care about. It is a strange thing to undervalue, because it is the foundation of our political system, but it would also be strange if a mild offense against the system canceled out all other political issues. We should consider what the optimum percentage should be—the number sufficient to keep politicians from body-slamming journalists. “If they anticipate that pushback, they will not even try,” Svolik said. Evidently 3.5 percent is not enough deterrence, because Gianforte went berserk on Ben Jacobs. But maybe 10 percent would have been enough. Think of these democracy-loving voters as antibodies: You don’t need a bloodstream brimming with antibodies to keep a disease at bay; you need a certain number, and anything more is superfluous. Ninety percent would be excessive. But at 3.5 percent, you get violence against journalists, and voter-ID laws that discriminate against minorities.

Finally, according to Graham and Svolik’s surveys, people are remarkably good at identifying undemocratic behavior as undemocratic. No matter the state of democratic decay, ordinary people seem to understand what it means to destroy democracy. Turks know that stopping your opponents from campaigning is an attack on democracy. Venezuelans know that shutting down a newspaper is an attack on democracy. And Americans know too. This at least suggests that some vestigial understanding of the principles of just government remains, even if just government is getting bargained down to fire-sale prices. The concept of democracy is not lost—and any concept that is still widely understood can be revived.