Earlier this month, the Democracy Fund Voter Study Group’s Lee Drutman released a fascinating study that challenged many of the dominant assumptions about the 2016 election. Drutman’s data suggested that Hillary Clinton and Bernie Sanders voters were largely aligned on economic issues, and that the 2016 election was decided by issues of culture and identity, rather than economics. Drutman concluded that Donald Trump’s victory largely stemmed from his ability to drive populists to the polls by hammering home the importance of protecting America’s cultural identity and keeping immigrants out of the country.

The VSG study also explains the mystery behind the Obama-Trump voter—that odd figure who voted for Barack Obama in 2008 and 2012 and Donald Trump in 2016. According to Drutman’s studies, these were voters who simultaneously held prejudicial values and relatively liberal economic views. In 2008 and 2012, they were forced to prioritize their values and, without a candidate running an explicitly prejudicial campaign, largely came down on the side of the Democrats. In Donald Trump, however, these populists found a political unicorn: a candidate who shared both their right-leaning cultural values and their left-leaning economic ones.

I talked to Drutman about the lessons of the 2016 election, the realignment currently taking place in American politics, and how Trump might fare in 2020. This interview has been edited and condensed for clarity.

Your study found that, contrary to much of the coverage of both the election and the primaries, it was cultural issues, not economic ones, that ultimately decided the 2016 election.

If you compare the 2016 and the 2012 elections, 2016 was fundamentally about who should be an American and what it really means to be an American in ways that were off the table in 2012. Immigration was front and center. Whether Muslims should even be allowed into the United States was front and center. These were issues that were just not discussed in 2012—and discussing these issues really activated attitudes surrounding race and identity. There were a lot of voters who supported Obama who held these beliefs, but the campaign wasn’t about these issues in either 2008 or 2012.

By talking so much about the threat of immigration, and the threat of Muslim “infiltration” in America, Trump made these the most salient issues for many voters. In 2016 they decided that Trump was the candidate who understood how they felt about those issues, whereas Clinton was going to be an “open borders, letting all the Muslims in” candidate. That had an impact.

Did the Democrats’ shift on cultural issues help perpetuate this trend?

If you compare where Democrats were in 2008 to 2016, I think you would see that Democrats moved to the left on immigration. In a lot of places that had virtually no immigration for a very long time, people started to see more people speaking Spanish, or more people of Muslim heritage. If this is new to you—even if the Spanish-speaking population of your town goes from 1 percent to 3 percent and the absolute change is not high—the rate of change can feel very fast.

If you look at states that had the most rapid increases in immigration, it’s places like Arkansas. There still aren’t a lot of immigrants in Arkansas, but it feels like a huge shift. The Democratic Party also moved considerably left on these issues in a way that made it more challenging for some of these voters to support it. If Trump wanted to start a culture war, Clinton gave him the culture war he wanted to start.

Your study found that there wasn’t much difference between Clinton and Sanders voters on economic issues, though there were considerable differences in how they viewed the political establishment.

If you look at where Clinton voters placed themselves on economic issues and where Sanders voters place themselves on economic issues, there’s really no difference. There is a difference in how Clinton and Sanders talked about economic issues, of course. Sanders said he was going to take on the banks and take on the pharmaceutical companies. Hillary Clinton didn’t use that language. But there’s also a lot of evidence to suggest that a lot of Sanders’s support was motivated by the fact that his voters didn’t like Clinton and didn’t trust the Washington establishment. They felt that Clinton was an insider and that the process was set up to favor her. There’s something about the Clintons that a lot of Democratic voters don’t trust.

Do you think that the kind of economic arguments that Sanders made about money in politics and the growing power of corporations could be a way for Democrats to win back some of the populist voters they lost to Donald Trump?

As long as Donald Trump is the Republican standard bearer and pursues the policies he does on race and immigration, these identity issues are going to remain quite salient. It’s not clear to me that a more populist economic message will necessarily cut through given the loudness of the race and identity politics.

That said, I think the good news for Democrats on that front is twofold. One is that a lot of the people who voted for both Mitt Romney in 2012 and Hillary Clinton in 2016 also feel that economic inequality is a tremendous problem. They’re a little more skeptical about the government’s ability to do anything about it, but they share a lot of those concerns. Two is that there’s also a turnout and mobilization issue: it’s not just a question of reaching out to swing constituencies, it’s also a question of making sure your own constituencies turn out. As you know, midterm elections tend to have much lower turnout than presidential elections.

If Democrats have bold economic policy ideas that excite their core voters, that’s a way of getting voters to turn out and feel like they’re turning out for something. A lot of them will turn out just because they want to beat Trump, but perhaps not everybody. The younger you go in the electorate the more dominant that liberal quadrant is, whereas the populist quadrant is mostly older.

Reading your study I couldn’t help but come to the conclusion that both parties are in the middle of an existential crisis. Do you think we’re in the middle of a party realignment right now?

I think that’s right. I was writing last year that Trump heralded a realignment, in the sense that he has caught up to where the Republican electorate has been drifting for a while. What you’ve seen over several decades is a shift in which the Republican Party has become the party of the white working class, which used to predominantly vote Democrat. This shift started with Reagan Democrats and it has continued ever since. Meanwhile, the Democrats have increasingly become the party of the professional class and ethnic minorities.

Parties take a while to catch up to their voters, because party leaders tend to give up power reluctantly—they want to hold on to how things used to be because that keeps them in power. The struggle with Trump transitioning the Republican Party to being a more populist, more ethnographic-nationalist party is that there is no intellectual infrastructure or policy infrastructure to implement the vision that Trump ran on. He certainly doesn’t offer it himself. What you’re starting to see now is how some folks on the right are wrestling with what it means to put forward a vision that fits in this nationalist, populist framework. As that vision expands to eventually take over the party, what you’ll also see is an exodus from the party of more traditionalist Republicans.

Do you see continued friction between this older party infrastructure and Trump’s electorate?

The Republican Party is very much a party in transition now, as it’s caught between the remnants of an older policy vision that no longer fits and the electorate that it now represents. Presumably there’s going to be some backlash when voters who were excited about Trump in the primary—precisely because he filled a void within the Republican Party—see him put forward policies that are antithetical to the policies that he promised. Maybe he can get away with that because these voters have rejected the Democratic Party on cultural and identity grounds, but I suspect they will continue to want the Republican Party to be the party that Trump promised it would become.

I think you will start to see more candidates, like [the neo-Confederate] Corey Stewart in Virginia, coming forth. One possible reason for why Stewart came close to beating Ed Gillespie in the Republican primary for governor was that a lot of people who would have been Gillespie supporters eight years ago no longer consider themselves Republicans, and they now vote for Democrats.

The fault lines that Trump exposed in 2016 aren’t going anywhere. Even with Democrats confident that the unpopularity of American Health Care Act will propel them to victories in 2018 and 2020, Trump seems to be set on hitting immigration and identity over and over again.

Hitting immigration and identity is what brought him to the White House. If that’s the most salient issue for voters, they will stay supporters no matter what he does because he’s picked the right enemies and he’s signalled that he’s on their side. If the focus is on whether he’s still hanging on to his promises of delivering government entitlements, he loses. If the question is over American identity, he has a chance of retaining support.

If you think about it, the takeaway—which is a broad takeaway on American politics—generally on economic issues, on welfare issues, the country’s overwhelmingly liberal. On cultural issues, the American population is conservative, particularly in rural areas that tend to be overrepresented in our system of government.

Trump was able to pitch himself to voters as a marriage of social conservatism and economic liberalism, arguing that he would maintain (and possibly even expand) the welfare state. He’s clearly going back on many of the promises he made on the campaign trail and he won’t be able to run as a unicorn in 2020. Could that cost him his presidency?

Well, it may. That may be the case, but it may also be the case that as long as Trump maintains the right enemies, then he seems like he’s still the lesser of two evils to many voters. Republicans may succeed by making Democrats look like the party of globalist multiculturalism undermining American Christian greatness. That could still be enough to win.