Why did the 2016 election look so much like the 2012 election?

Javier Zarracina/Vox

A few months ago, I stopped by Larry Bartels’s office at Vanderbilt University. Bartels, alongside Christopher Achen, is the author of Democracy for Realists, which I’d become a bit obsessed with. The book argues that decades of social science evidence has shattered the idealistic case made for how voters in democracies act, and the reality is that “even the most informed voters typically make choices not on the basis of policy preferences or ideology, but on the basis of who they are — their social identities.”

I sat down with Bartels shortly after the 2016 election, and I had a dozen ideas for how his book helped explain the unusual results. But he wasn’t buying my premise. To him, the election looked pretty typical.

The Democratic candidate won 89 percent of Democratic voters, and the Republican candidate won 90 percent of Republican voters. The Democrat won minorities, women, and the young; the Republican won whites, men, and the old. The Democrat won a few percentage points more of the two-party vote than the Republican, just as had happened four years before, and four years before that. If you had known nothing about the candidates or conditions in the 2016 election but had been asked to predict the results, these might well have been the results you’d predicted. So what was there to explain?

Bartels doesn’t deny that there were interesting oddities to this election. The small but crucial number of Obama-to-Trump voters are worth studying, for instance. The interventions of Russia and then-FBI Director James Comey may well have delivered Republicans the presidency. And surely it’s meaningful that Hillary Clinton and Donald Trump were the least popular nominees in history.

But his point is that we’re so obsessed by what was different in 2016 that we’re missing the big story — how much stayed the same. For all the weirdness of the campaign, Trump and Clinton still got about 95 percent of the vote, and they did so by consolidating their own bases in ways that looked extremely similar to 2012. It’s easy to come up with counterfactuals where Clinton is in the White House today, or where Marco Rubio won the popular vote by 4 percentage points, but the basic similarity between Trump-Clinton and Obama-Romney deserves more attention than it’s gotten.

If democracy is a contest of ideas, if it’s really about voters judging candidates and policies and making fresh judgments about the world around them, how come a wild-card nominee wielding a brand new ideology didn’t do more to shuffle the deck? How did American politics become so stable that even an election as weird as 2016 ended up looking normal? And if it’s not a contest of ideas, then what the hell is it?

In Bartels’s view, there is an answer. He thinks this election told us something we already knew. The problem is that it’s something nobody wants to hear.

The problems with democracy

Bartels and Achen describe Democracy for Realists as “a kind of intellectual conversion experience for us.” It had long been obvious to them, as it is to everyone, that the practice of democracy is often a grubbier, shabbier thing than the glittering rhetoric that surrounds it. But that’s no great crime. “We believed that if the realities failed to match the ideals, we (and others seeking to vindicate contemporary democracy) still had intellectually powerful back-up defenses to bolster our convictions,” they write.

And then, as they surveyed the evidence, it turned out they didn’t. “Our view is that conventional thinking about democracy has collapsed in the face of modern social-scientific research.”

Most of Democracy for Realists is Achen and Bartels systematically, ruthlessly demolishing traditional theories of democracy. Neither their arguments nor their evidence is surprising, at least not exactly. What’s unusual is their willingness to admit what the research actually reveals. This book is the political science equivalent of being told Santa doesn’t exist: It makes sense once you think about it, but everyone has spent years telling you otherwise, and so the revelation comes as a trauma.

Achen and Bartels present the main argument for democracy, the one we all learned in grade school, like this: “Ordinary people have preferences about what their government should do. They choose leaders who will do those things, or they enact their preferences directly in referendums. In either case, what the majority wants becomes government policy.” The authors call this the “folk theory” of democracy, and it is a persuasive, inspiring vision of how government should work.

But political scientists rejected that argument decades ago, and much of Democracy for Realists consists of Achen and Bartels running through the depressing studies that explain why. Voters hold weak and contradictory views about the government. They pay little attention to political news save in the months directly before an election. Their positions on key issues can be changed by small tweaks to how questions are worded or ordered. They often assume the leaders they like share their views, or if they don’t, they change their views to accord with the politicians they support. Look no further than the sudden rise in Republican estimations of Vladimir Putin to see this phenomenon in action:

None of this means voters are stupid. Quite the opposite, actually. The US government deals with a dizzying range of questions, and developing informed and current opinions on all of them would take more hours than any of us have in a day. Meanwhile, most people have jobs to do, families to feed, friends to see, lives to live, and very little actual political power to wield — so they sensibly don’t spend their scarce free time developing detailed views about appropriations bills and trade deals and conservation policy and military spending and corporate tax reform and Medicaid funding and the proper path of interest rates.

But if voters aren’t judging candidates based on their preferences about what government should do, then how are they making decisions? How, to put it bluntly, does democracy actually work?

In recent decades, the most academically influential model has been the “retrospective theory of voting.” This theory holds that voters use assessments of current conditions — like looking at whether the economy is growing, or if Americans are dying in an overseas war — to decide whether the party in power is doing a good job. In this telling, voters are taking a simple but powerful shortcut to the place that deeply informed opinions about public policy would get them: a country that runs well, an economy that grows quickly, a world that is peaceful.

There is nothing new about this theory of how voters actually act. “To support the Ins when things are going well; to support the Outs when they seem to be going badly,” wrote Walter Lippmann in 1925, “this, in spite of all that has been said about tweedledum and tweedledee, is the essence of popular government.”

The retrospective theory of voting has the advantage of being true. Economic growth really does drive election results, for instance. But it’s true in a haphazard, problematic way. Studies find that the only economic growth that actually matters is the growth that happens in an election year, and there’s no evidence that voters can separate a recession a president had nothing to do with, and perhaps even managed well, and a recession a president’s policies caused or worsened.

“If jobs have been lost in a recession, something is wrong, but is that the president’s fault?” write Achen and Bartels. “If it is not, then voting on the basis of economic conditions may be no more sensible than kicking the dog after a hard day at work.”

A yet clearer example comes from the well-established finding that voters punish incumbents for bad weather and natural disaster. Here, Achen and Bartels are at their acidic best:

The fact that American voters throughout the 20th century punished incumbent presidents at the polls for droughts and floods seems to us to rule out the possibility that they were reacting to subpar handling of misfortunes rather than to the misfortunes themselves. After all, it is hard to see how incumbent presidents’ handling of droughts and floods could have been substantially worse than average over the course of an entire century.

One problem with the retrospective theory of voting, in other words, is that the shortcuts get us lost. “Voters consistently and systematically punish incumbents for conditions beyond their control,” conclude Achen and Bartels.

But perhaps a bigger problem for the theory — at least as an explanation of voter behavior — is that most voters don’t use it at all. Particularly in modern elections, the swingable vote is a tiny fraction of the population. In 2008, for instance, the Republican Party was presiding over an economic meltdown and an unpopular war, but John McCain still got 46 percent of the vote. Most voters support their party’s nominee for the presidency no matter the condition of the country.

Let’s recap. The classic argument for democracy is that people have clear preferences for what government should do and they vote in accordance with those preferences. That’s not true. The more modern, and more academically influential, argument for democracy is that voters make judgments about how incumbents perform, and those judgments are broadly accurate and work as a shortcut for getting them the kinds of leaders they want. That is also not true.

“All the conventional defenses of democratic government are at odds with demonstrable, centrally important facts of political life,” write Achen and Bartels. “One has to believe six impossible things before breakfast to take real comfort in any of them.”

So how, then, do voters make decisions in a democracy?

It’s the identity, stupid

During the worst of Northern Ireland’s “Troubles,” when tensions between Catholics and Protestants were at their height, the Irish poet Seamus Heaney told of a visitor to the region who was asked whether he was Protestant or Catholic. The man replied that he was an atheist.

“Yes, yes, we understand,” his hosts said. “But are you a Protestant atheist or a Catholic atheist?”

Donald Trump might have been, in political if not religious terms, an atheist, but he was clearly a Republican atheist. He knew which side he was on, even if he didn’t believe what they believed. And that was enough. That was more than enough. Tribes exist to fight common enemies.

Achen and Bartels believe we’ve spent so much time listening to what voters say that we’ve lost sight of what they actually do, and why they do it. “In thinking about politics,” they write, “it makes no sense to start from issue positions — they are generally derivative from something else. And that something else is identity.”

This is a profound statement, and the authors don’t shy away from its scope. “A realistic theory of democracy must be built, not on the French Enlightenment, on British liberalism, or on American Progressivism, with their devotion to human rationality and monadic individualism, but instead on the insights of the critics of these traditions, who recognized that human life is group life,” they write.

In this telling, democracy is less a contest between ideas than a contest between identities, with both parties constantly trying to activate the basket of identities that will lead to a vote for their tribe. Are you a Republican or a Democrat? Urban or rural? Catholic or Jewish? White or black? Male or female? Liberal or conservative? Rich or poor?

Even gentle, isolated reminders of identity can swing people’s voting behavior. Take a 2014 study by psychologists Maureen Craig and Jennifer Richeson. They asked one group of white political independents if they knew that California would soon have more nonwhite residents than white residents. Then they asked another group of white political independents if they had heard another highly racialized, but less threatening (at least to white political power), fact — “that Hispanics had become roughly equal in number to Blacks nationally.” And then they compared the two group’s political opinions.

Even now, the results, as described by Achen and Bartels, stun. “The people who had been informed (or simply reminded) of the potentially threatening demographic shift in California were significantly more likely to lean Republican. This effect was twice as strong in the West as in the nation as a whole, producing a substantial 11-point increase in Republican leaning (and a 15-point decrease in Democratic leaning).”

In a follow-up study, Craig and Richeson handed some white subjects a press release about “projections that racial minorities will constitute a majority of the U.S. populace by 2042.” The group that read the release “produced more conservative views not only on plausibly relevant issues like immigration and affirmative action, but also on seemingly unrelated issues like defense spending and health care reform.”

It’s not just white voters whose political opinions are easily swayed by simply contemplating threats to their racial group’s status and identity. A 2016 study by Alexander Kuo, Neil Malhotra, and Cecilia Hyunjung Mo split a sample of Asian-American college students into two groups. One group was subjected to a staged racial humiliation during the study — their US citizenship was doubted by the researcher managing the experiment. “This minor but socially charged interaction boosted Democratic partisanship by 13 percentage points,” report Achen and Bartels.

It is worth dwelling on what isn’t happening here: any persuasion about policies or ideas. These experimental cues trigger complex, multilayered identities. They do not make arguments about ideal marginal tax rates or America’s role in the world. Results like these do not make sense under the traditional or retrospective theories of democracy; they make perfect sense under the group identity theory of democracy.

When we talk about group identities in American politics, we mainly mean nonpolitical identities: race, class, gender, sexuality. But there’s one identity that elections activate above all, and that identity is becoming more and more powerful: partisan identity.

One identity to rule them all

There is one overwhelming fact that structures American politics, and it is this: People who vote for Republicans vote for Republicans, and people who vote for Democrats vote for Democrats. It might sound tautological, but it isn’t. A few decades ago, people who voted for Republicans often voted for Democrats, and vice versa. Split-ticket voting was common, and even hardcore, self-described partisans were often persuadable.

Not anymore. There are a few findings that rocked my understanding of politics, and one of them came from political scientist Corwin Smidt. Looking at decades of election data, he found that self-described independent voters today are more loyal to a single party than voters who described themselves as “strong partisans” were in the 1970s. This bears repeating: The people who say they’re free from either party today are more partisan in their voting habits than the people who said they were strong loyalists of a single party in the ’70s.

Smidt argues that the change is in our parties, not in our voters. The difference between the Democratic and Republican parties has become so clear, so unmistakable, that pretty much every kind of voter reliably votes for one party or the other. In 1964, Medicare — a single-payer health care system for the elderly — received 70 Republican votes in the House as well as 13 Republican votes in the Senate. There was no anti-abortion plank in the GOP platform until 1980. It's easy to see how a voter in the 1970s might think Republicans were open to something like Medicare or reproductive choice — particularly if they lived in a liberal area represented by a liberal Republican who actually was open to those policies.

Today, however, the choice between the two parties is much, much clearer. You may not like Donald Trump, but you fear Hillary Clinton. As the parties diverge from each other ideologically and culturally, the other side becomes more of a threat — and that makes it easier to justify voting for your side, no matter who the nominee is.

Polling backs this up. Since 1964, the American National Election Studies have been asking Republicans and Democrats to describe their feelings toward the other party on a scale that runs from cold and negative to warm and positive. In 1964, 31 percent of Republicans had cold, negative feelings toward the Democratic Party, and 32 percent of Democrats had cold, negative feelings toward the Republican Party. By 2012, that had risen to 77 percent of Republicans and 78 percent of Democrats. Today, fully 45 percent of Republicans, and 41 percent of Democrats, believe the other party’s policies “threaten the nation’s well-being.” Fear of the other is a powerful force to keeping unity in a tribe.

But policy does not account for the entirety of the rise in political tribalism. In the New York Times, Amanda Taub ran through new research suggesting that “party affiliation has become an all-encompassing identity that outweighs the details of specific policies.” The key idea here is what we might call identity amplification, which has become more prevalent in recent years as people have clustered themselves economically, ideologically, religiously, and geographically:

Everyone has multiple identities: racial, religious, professional, ideological and more. But while those multiple identities might once have pushed people in different partisan directions — think of the conservative Democrats of old in the South or all the liberal Republicans in the Northeast — today it’s more common to line up behind one party. A white conservative who lives in a rural area and is an evangelical Christian is likely to feel that the Republican Party is the best representative of all of those separate identities, for instance. An African-American liberal who lives in a city and works in a professional job is likely to feel the same way about the Democratic Party.

Identity amplification makes Republican and Democratic identities overwhelmingly powerful, and perhaps even inescapable, in modern politics — and modern life.

In 1960, Americans were asked whether they would be pleased, displeased, or unmoved if their son or daughter married a member of the other political party. Then, only 5 percent of Republicans, and only 4 percent of Democrats, said they would be upset by the cross-party union. Fast-forward to 2008. The polling firm YouGov asked Democrats and Republicans the same question — and got very different results. This time, 27 percent of Republicans, and 20 percent of Democrats, said they would be upset if their son or daughter married a member of the opposite party. In 2010, YouGov asked the question again; this time, 49 percent of Republicans, and 33 percent of Democrats, professed concern at interparty marriage.

Today, studies find that people are more willing to discriminate against someone of the other party than of another race, at least in experimental settings. "Political identity is fair game for hatred," Shanto Iyengar, director of Stanford's political communications lab, told me. "Racial identity is not. Gender identity is not. You cannot express negative sentiments about social groups in this day and age. But political identities are not protected by these constraints. A Republican is someone who chooses to be Republican, so I can say whatever I want about them."

This, Achen and Bartels say, is what elections are really about; “political campaigns consist in large part of reminding voters of their partisan identities.” All those ads, those speeches, those conventions, those endorsements, those photo ops — they’re not about persuading voters of whose tax plan is better so much as they’re about reminding voters which tribe they’re from.

Trump understood all this, even if he didn’t embody it — he is a political independent, a lifelong Manhattanite, and a billionaire libertine who spent the campaign affirming evangelical, rural, working-class identities. He routinely contradicted or dismissed longtime Republican ideas, but what he never, ever did was disrespect Republican identities. Liberals laughed when Trump said, “I love the poorly educated," but his tribe knew what he meant — unlike those smug liberals, he was on their side and thought they deserved more respect.

Viewed this way, Trump’s communication style makes more sense. His extemporaneous, rambling, quasi-factual speeches confuse pundits — including me — who are used to hearing politicians make careful arguments. But his speeches do work to establish which side he’s on, which groups he admires, and which enemies he’s going to humiliate. Perhaps more effectively, his tweets and insults drive his opponents into frenzies, and make the battle lines clear — you may not like Trump, but if you’ve spent years hating the Democrats and the media and condescending professors and rich cultural elites, at least he’s pissing them off and, in doing so, proving he’s on your side.

And even if you didn’t like Trump — and many Republicans didn’t — fear and mistrust of the other side was a perfectly rational reason to vote for him. There was a Supreme Court seat up for grabs, after all, and conservatives reasonably feared whom Clinton would appoint. This is one way polarization amplifies identity: The two parties are now so far apart ideologically that even an unusual nominee like Trump is a safer bet than the Democrat. Many Trump voters appear to have made this calculation. According to exit polls, Trump won among voters who said their vote was motivated by dislike for the other candidates in the race.

This is the answer to the normalcy of the 2016 election: As abnormal and unqualified and erratic as Trump showed himself to be, with Clinton’s help, he still activated pro-Republican and anti-Democratic identities, and that left him a stone’s throw, or a Jim Comey letter, away from the presidency. Yes, there were a few Obama-to-Trump voters on the margins, but overwhelmingly, even an election this weird was forced into normalcy by the power of our partisan identities.

Back to top ↑