(Paul J. Richards/AFP/Getty Images)

In the year since Donald Trump was elected president of the United States, many have continued to debate what motivated those who voted for him. We wanted to understand Trump voters among millennials, the generation that will make up the largest share of the voting-eligible population in the upcoming 2018 midterms. Most millennials voted for Clinton, as predicted.

But when we disaggregate the millennial vote by race and ethnicity we find some interesting things — including, notably, that 41 percent of white millennials voted for Donald Trump in 2016.

Why?

Two popular explanations have emerged post-election: 1) economic anxiety, and 2) racial resentment. Many commentators have argued that a sense of economic loss drove many white working-class voters toward Trump. Meanwhile, here at the Monkey Cage, Michael Tesler has explained that support for Trump was especially linked to racial resentment.

Examining our data from the GenForward Survey, we find a hybrid explanation. First, white millennial Trump voters were likely to believe in something we call “white vulnerability” —  the perception that whites, through no fault of their own, are losing ground to other groups. Second, racial resentment was the primary driver of white vulnerability — even when accounting for income, education level or employment.

Here’s how we did our research

Our data come from a GenForward Survey fielded Aug. 31 to Sept. 16, which included 1,816 respondents between 18 and 34. The GenForward Survey is a bimonthly, nationally representative survey of young adults, with particular focus on how race and ethnicity shape political attitudes. This sample includes responses from 503 African Americans, 510 whites, 505 Latinos and 258 Asian Americans, which are weighted to population estimates and allow us to compare group attitudes with a higher degree of confidence than most other surveys of this age group.

To measure white vulnerability, we constructed a scale out of three questions, asking: 1) whether whites were “economically losing ground through no fault of their own”; 2) whether discrimination against whites was “as big a problem as that against Blacks and other minorities”; and 3) if minorities overtaking whites as the majority of the U.S. population by 2050 would “strengthen or weaken the country.”

To assess whether and how this sense of white vulnerability was related to racial attitudes, we used the standard racial resentment scale. Racial resentment is characterized by “a moral feeling that blacks violate such traditional American values as individualism and self-reliance” and has been linked to a number of political attitudes.

To examine other potential explanations, we relied on standard measures of education, employment, income, partisanship, ideology, economic perceptions, gender, employment, marital status, residential context and region.

Here’s what we found

As we mentioned above, our data show that 41 percent of white millennials voted for Trump in 2016, an estimate that largely mirrors national exit polls. About 84 percent of millennial Trump voters were white. Compared to white voters who did not support Trump, Trump voters were more likely to be male, married  and without college education. Other possible differences — like geographic region and living in a metropolitan area — were negligible between white Trump and non-Trump voters.

When we control for a number of other factors that might lead white millennials to vote for Donald Trump — such as racial resentment, partisanship, ideology, living in the South, gender and employment status — we find that the largest predictor of voting for Trump is that sense of white vulnerability. White millennials who scored high on the white vulnerability scale were 74 percent more likely to vote for Trump than those at the bottom of the scale.


What leads to “white vulnerability”? 

So what’s behind feelings of white vulnerability? Not going to college is one predictor. If we’re looking just at white millennials who scored high on the white vulnerability scale compared to those who did not, the probability of voting for Trump among those who have not gone to college was 95 percent. For those who have gone to college, by contrast, the probability of voting for Trump was only 28 percent.

In other words, a white millennial with a high school education and strong perception that whites are losing ground to other groups through no fault of their own was almost certainly a Trump voter; less than one in three of those who went to college and held similar perceptions of white vulnerability were Trump voters.

So are young Trump voters the white working-class folks who felt economically vulnerable, as so many observers have proposed? That’s not exactly what we found.

Contrary to what some have suggested, white millennial Trump voters were not in more economically precarious situations than non-Trump voters. Fully 86 percent of them reported being employed, a rate similar to non-Trump voters; and they were 14 percent less likely to be low income than white voters who did not support Trump. Employment and income were not significantly related to that sense of white vulnerability.

So what was? Racial resentment.

Even when controlling for partisanship, ideology, region  and a host of other factors, white millennials fit Michael Tesler’s analysis, explored here. As he put it, economic anxiety isn’t driving racial resentment; rather, racial resentment is driving economic anxiety. We found, as he has in a larger population, that racial resentment is the biggest predictor of white vulnerability among white millennials. Economic variables like education, income  and employment made a negligible difference.

But when white millennials scored high on racial resentment they were 42 percentage points more likely to indicate feelings of vulnerability than those who scored low — and therefore much more likely to vote for Trump.

Other analysts have found something similar, including Emma Green at the Atlantic, Jennifer Rubin here at The Washington Post  and Michael Tesler and John Sides here at the Monkey Cage — all of whom suggested that those voting for Trump felt what we would call white vulnerability, racial resentment, and mixed in with that, an anxiety about losing cultural status.

And as Ben Casselman wrote at FiveThirtyEight, anxiety is “all about what lies ahead.” Anxiety about cultural slippage may include fear of economic loss. But it’s not the same as actual economic fragility.

Many white Americans are uneasy with what they see as their future, surrounded as they are by growing racial and cultural diversity in mainstream media, politics, entertainment and music. White millennials are part of the U.S.’s most diverse generation, as so many have discussed — but not all of them are comfortable with it. Voting for Trump reveals those racial and cultural  anxieties.

Matthew Fowler and Vladimir E. Medenica are postdoctoral scholars for the GenForward Survey at the University of Chicago.

Cathy Cohen is the founder/principal investigator of the GenForward Survey and the David and Mary Winton Green professor of political science at the University of Chicago.