The Doom Spiral of Pernicious Polarization

The U.S. is more dangerously divided than any other wealthy democracy. Is there a way back from the brink?

Red and blue magnetic poles
The Atlantic; Getty

Until a few decades ago, most Democrats did not hate Republicans, and most Republicans did not hate Democrats. Very few Americans thought the policies of the other side were a threat to the country or worried about their child marrying a spouse who belonged to a different political party.

All of that has changed. A 2016 survey found that 60 percent of Democrats and 63 percent of Republicans would now balk at their child’s marrying a supporter of a different political party. In the run-up to the 2020 presidential election, the Pew Research Center reported that roughly nine out of 10 supporters of Joe Biden and of Donald Trump alike were convinced that a victory by their opponent would cause “lasting harm” to the United States.

As someone who lived in many countries—including Germany, Italy, France, and the United Kingdom—before coming to the United States, I have long had the sense that American levels of partisan animosity were exceptionally high. Although I’d seen animosity between left and right in other nations, their hatred never felt so personal or intense as in the U.S.

A study just published by the Carnegie Endowment for International Peace confirms that impression. Drawing on the Variety of Democracies (V-Dem) data set, published by an independent research institute in Sweden that covers 202 countries and goes back more than two centuries, its authors assess to what degree each country suffers from “pernicious” levels of partisan polarization. Do their citizens have such hostile views of opponents that they’re willing to act in ways that put democracy itself at risk?

The authors’ conclusion is startling: No established democracy in recent history has been as deeply polarized as the U.S. “For the United States,” Jennifer McCoy, the lead author of the study and a political-science professor at Georgia State University, told me in an interview, “I am very pessimistic.”

On virtually every continent, supporters of rival political camps are more likely to interact in hostile ways than they did a few decades ago. According to the Carnegie study, “us versus them polarization” has been increasing since 2005. McCoy and her colleagues don’t try to explain the causes, though the rise of social media is obviously a contributing factor.

As near-universal as political polarization has become, it is more pronounced in some places than in others. On a five-point scale, with 0 indicating a country with very little partisan polarization and 4 indicating a country with extreme polarization, both the U.S. and the rest of the world, on average, displayed only a modest degree of polarization at the turn of the millennium: They each scored about a 2.0. By 2020, the world average had increased significantly, to a score of about 2.4. But in the United States, polarization accelerated much more sharply, growing to a score of 3.8.

Among countries whose political institutions have been relatively stable, the pace and extent of American polarization is an eye-popping outlier. “Very few countries classified as full liberal democracies have ever reached pernicious levels,” the study’s authors write. “The United States stands out today as the only wealthy Western democracy with persistent levels of pernicious polarization.” When I spoke by phone with McCoy, she was even more categorical: “The situation of the United States is unique.”

To live in a country where political disagreements turn into personal vendettas is no fun, but a growing body of research reveals more systemic effects. Pernicious polarization makes good-faith efforts to tackle social problems such as public-health crises harder and bad-faith efforts to turn them into political gain easier. At worst, an erosion of trust in democratic norms and political institutions can end up as political violence and civil war.

The fundamental premise of democracy is that citizens agree to be ruled by whoever wins an election. But if many citizens come to believe that letting the other side rule poses a threat to their well-being, even their lives, they may no longer be willing to accept the outcome of an election they lose. This makes it easier for demagogues to attract fervent supporters, and even to turn them against a country’s political institutions. The January 6, 2021, assault on the Capitol is just such a symptom of the malaise.

We might hope that the history of other nations would offer clues to how the U.S. could get its polarization under control. The past century has yielded notable cases of “depolarization,” from Italy in the 1980s to Rwanda in the early 2000s. In Italy, escalating political violence from both the far left and the far right had threatened to tear the country apart, but leaders from rival political parties eventually united against terrorism, and that enabled the country to weather the crisis. In 1990s Rwanda, Hutus murdered hundreds of thousands of Tutsis in an orgy of genocidal violence, yet the country has achieved a modicum of national reconciliation and managed to keep the peace (though at the price of Paul Kagame’s autocratic leadership). Do examples like these contain any useful lessons for the U.S.?

Unfortunately, the data in the Carnegie study do not offer much cause for optimism. About half of the time a country experienced serious polarization since 1900, mutual distrust and hatred turned into a permanent condition. Although political tensions waxed and waned, these countries never fell below the level of pernicious polarization for any extended period. And many countries never recovered. Once pernicious polarization has set in, it stays.

That leaves the other half of cases. Those don’t offer much hope, either. Many of the supposed success stories saw either a relapse into dangerous levels of polarization or merely a moderate degree of depolarization. And when a country did manage to depolarize in a lasting way, a major political disaster seemed needed to force it: a civil war, a cruel dictatorship, or a struggle for independence. Only after overcoming such dire turmoil did these countries escape their vicious cycles. “The prevalence of systemic shocks in bringing about depolarization,” the study’s authors note, “was especially striking.”

That no such systemic shock has struck the U.S. in modern history would seem to bode ill for American prospects of depolarization. But do things really have to fall apart before we can put them back together?

The state of America’s union is especially fractious, true, but our predicament may not be quite as dire as it seems. The limitations of the Carnegie study itself illustrate why we should take predictions of doom with a grain of salt.

The survey’s polarization data are ambitious in scope—aggregating 120 years of historical information about a large number of countries—but the methodology behind them is more modest. The V-Dem data set used in the Carnegie study relies on asking a group of from five to seven country experts a single question about any given nation: “To what extent is society divided into mutually antagonistic camps in which political differences affect social relationships beyond political discussions?” If the experts answer that this is the case to a “noticeable extent,” with supporters of opposing camps “more likely to interact in a hostile than friendly manner,” this counts as a 3, on a scale of 0 to 4. That score is enough to qualify as “pernicious polarization.” What’s more, this assessment is highly retrospective: How polarized America was in, say, 1935, or in 1968, or in 1999 is a judgment made by a handful of social scientists only recently. Quantifying polarization like this is susceptible to distortion in two ways: presentism and provincialism.

Experts evaluating how polarized America has been in the past century might remember every detail of a shouting match with a Trumpy uncle at last year’s Thanksgiving, but they cannot possibly have such a visceral feel for political divisions in, say, the 1910s, however much they’ve read about the period. That risks presentism. With their personal experience of partisan conflict and the shrill tone on social media top of mind, they may overestimate how much partisan hatred there is today and underestimate how much partisan hatred there was in the past.

“Expert surveys are subjective,” McCoy admitted when I put this concern to her. “There is no way of getting around that.”

What’s more, these experts will surely have different cultural assumptions about what constitutes a hostile political interaction. This—the danger of provincialism—makes comparisons among countries more difficult. In America, what is salient is how much nastier and more aggressive political discourse has become in recent decades. But in a society that recently experienced civil war, what may be more salient is that people are now willing to disagree about politics without killing one another. That perceptual gap between what counts here and what counts there could lead experts to assess a comparatively peaceful country that has become more sharply divided, like the U.S., as more polarized than a war-ravaged country that is somewhat less divided than it used to be.

In many of the countries that have experienced pernicious polarization, partisan political identity aligns almost perfectly with visible markers of ethnic or religious identity. In Lebanon and Kenya, for example, it is enough to see or hear a person’s name to know which way they’re likely to vote. When polarization spikes in these places, supporters and opponents of a political candidate don’t just yell at each other; they withdraw from all social cooperation, and their animosity grows vengeful and deadly.

“If you work for the Croat Catholic fire department,” Eboo Patel, a prominent interfaith leader, writes about Mostar, in Bosnia and Herzegovina, “you don’t respond to the burning buildings of Bosnian Muslims, even if you happen to be closer. And if you work for the Bosnian Muslim fire department, you let the flames engulf Croat Catholic homes.”

America’s polarization clearly differs from the Bosnian example—which experts actually scored as significantly lower than the United States’ (at 3.2 out of 4)—in two crucial ways. First, the overlap between partisan polarization and divisions of race, class, or religion is at best imperfect in the U.S. Although demographic patterns do offer clues to the likelihood of people’s support for Democrats or Republicans, a significant number of Latino Americans vote for the Republican Party and a lot of white Americans vote for the Democratic Party. Second, in many spheres of American life—including the workplace and Little League games—people put aside political differences or may not even be aware of them. And the local fire department does not ask for your voter registration before deciding whether to put out your house fire.

Perhaps America is not so much uniquely polarized as polarized in a unique way. Fifty years ago, out-group hatred in the United States primarily involved race and religion: Protestants against Catholics, Christians against Jews, and, of course, white people against Black people. Most Americans did not care whether their children married someone from a different political party, but they were horrified to learn that their child was planning to “marry out.”

Today, the number of Americans who oppose interracial marriage has fallen from well over nine in 10 in 1960 to far less than one in 10. And as the rapid increase in the number of interracial babies shows, this is not just a matter of people’s telling pollsters what they want to hear. In contrast to the dynamic in other deeply polarized societies, the division in America between opposing political camps revolves less around demographics and more around ideology.

A host of recent social-science studies backs this up. In one experiment, a sample of Americans was asked to award scholarships to fictitious high-school students. Presented with a candidate’s résumé suggesting that the applicant was of a different racial group from theirs, the subjects engaged in surprisingly little discrimination. (In fact, Americans of European descent tended to favor, not discriminate against, African American candidates.) But presented with a résumé that suggested the applicant had a different political-party affiliation from theirs, the subjects had a strong tendency toward bias: When choosing between similarly qualified scholarship candidates, four out of five Democrats and Republicans favored an applicant who belonged to the same political party.

Even as American politics got nastier in recent years, the overlap between ethnic identity and partisan polarization has actually continued to weaken. Trump was competitive in the 2020 election in part because he significantly increased his 2016 share of the vote among Black, Asian-American, and especially Latino voters. And Biden won appreciably more white voters than Hillary Clinton did. In other words, a voter’s racial identity was much less predictive of voting behavior in 2020 than it was in 2016.

Jennifer McCoy affirmed this, when I asked her about the difference between the United States and other perniciously polarized democracies: “Unlike many other polarized democracies, we are not a tribal country based on ethnicity … The key identity is party, not race or religion.”

America’s uniqueness could allow a more hopeful story than the headline findings of the Carnegie report might suggest. If polarization is mainly a matter of partisan political identities, the problem may be easier to solve than divisions based on ethnic or religious sectarianism.

One approach that could alleviate polarization in the U.S. is institutional reform. Right now, many congressional districts are gerrymandered, shielding incumbents from competitive primaries while making them hostage to the extremist portion of their base. Some states have attenuated this problem by taking districting out of party control. But other measures, such as adopting the single transferable vote or creating multimember districts, could also shift political incentives away from polarization.

California has already adopted a small-seeming—and thus realistic—innovation. In so-called jungle primaries, candidates from all parties compete in the election’s first round; then the top two finishers face off in the second-round general election. As a result, moderates with cross-party appeal get a fighting chance at being elected. If this can work in deep-blue states like California, it can work in deep-red states like Alabama.

Soothsayers of doom are in demand for a reason. American partisan polarization has, without a doubt, reached a perilous level. But America’s comparative competence at managing its ethnic and religious diversity, which has so far ensured that partisan political identities do not neatly map onto demographic ones, could be a saving grace.

We urgently need visionary leaders and institutional reforms that can lower the stakes of political competition. Imagining what a depolarization of American politics would look like is not too difficult. The only problem is that America’s political partisans may already hate one another too much to take the steps necessary to avoid catastrophe.