What explains the rise of Donald Trump?
There are many potential answers, but over the course of the campaign two competing theories have emerged. The first holds that Trump’s message appeals to working-class white voters who’ve seen their incomes stagnate, manufacturing jobs vanish, and inequality skyrocket in recent decades. The root cause of Trumpism, in this view, is economic insecurity. The other, blunter theory is that Trump’s fans flock to him for the same reason elites view him as an existential threat to American democracy: His open appeals to racist, white nationalist sentiment.
Both of these theories have some truth to them. But polling data suggests that racial resentment is the more important factor.
The American National Election Studies 2016 Pilot Study, a presidential primary extension of a long-running election survey, asked 1,200 eligible voters about the election, and their views on race, from Jan. 22 – 28, 2016. The poll had a number of questions designed to measure racial animus.
First, it asked respondents how important their race is to their identity. Second, it asked respondents whether they think the words “lazy” and “violent” describe black people, Muslims and Hispanics, “extremely well,” “very well,” “moderately well,” “slightly well” and “not well at all.” Finally, it included several questions meant to measure what scholars refer to as “racial resentment.” Developed by Donald Kinder and Lynn Sanders in 1996, the concept of racial resentment is designed to capture less overt, but still real, forms of racism. The concept is particularly useful for measuring racism these days, when most racism tends to be “colorblind” or “dog-whistle" racism -- that is, racist attitudes that are expressed in a way that is seemingly neutral, but still animates racial anger.
The survey then showed four prompts on racial resentment (which we then combined to a single metric), which each respondent was asked to rate on a 5-point scale from "strongly disagree" to "strongly agree":
- Irish, Italian, Jewish, and many other minorities overcame prejudice and worked their way up. Black people should do the same without any special favors.
- It’s really a matter of some people not trying hard enough; if black people would only try harder they could be just as well-off as whites.
- Over the past few years, black people have gotten less than they deserve.
- Generations of slavery and discrimination have created conditions that make it difficult for black people to work their way out of the lower class.
When analyzing this survey data, we threw in a number of statistical controls for individual race, age, income, education, partisan identification, political ideology, level of political interest, church attendance, perceptions of economic performance, and opinions about free trade and whether government should provide fewer or more services. That was meant to isolate the extent to which respondents’ views on race affected their views on the election.
On just about every measure, support for Trump increased along with the measured racial animus. As the chart below shows, increased levels of racial stereotyping among white respondents — as measured by belief that black people, Muslims and Hispanics are “lazy” or “violent” — strongly increases support for Trump, even after controlling for other factors. The opposite is true, however, when it comes to support for Marco Rubio. Among white respondents, support for Rubio decreases with belief in racial stereotypes:
The same story is true for racial resentment. The more troubling the respondents’ answers on the four resentment questions were, the likelier they were to support Trump. There is no such relationship between racial resentment and support for Marco Rubio or the other major Republican contenders:
While the racial stereotyping questions covered black people, Hispanics and Muslims alike, we also pulled out views about Muslims specifically to see how they affected Trump support. Once again, Trump support increases significantly among those who describe Muslims as “violent,” while the same does not hold for the other Republican candidates.
Other polling has found different results. A Vox/Morning Consult poll, for example, found that Rubio and Trump supporters were about equally likely to score high on racial resentment, with Rubio supporters potentially looking slightly worse. Dylan Matthews wrote, “Trump supporters were not unusually racially resentful compared with Rubio supporters.” But we think our analysis of the ANES survey is more reliable. Our analysis uses a model with controls, which helps isolate the effect of racism. In addition, the similarities between Trump and Rubio partially stem from collapsing together the “strongly agree”/“somewhat agree” and the “somewhat disagree”/ “strongly disagree” variables. When those categories are disaggregated, it becomes clear that Trump supporters are far more likely to strongly endorse racial resentment.
Worryingly for the GOP elite, our analysis also suggests one reason stopping Trump has proved so difficult: His support is highest among those who do not follow politics very closely. Support for Trump in January was strongest among those who said they followed politics “hardly at all,” while Kasich and Rubio performed best with those who follow politics more frequently. This means they may have missed, or simply don’t care about, many of the events (like the recent Chicago protests) that have pushed many away from Trump. In addition, Trump’s supporters are strongly anti-establishment, so even if they were aware of recent events like Mitt Romney’s denunciation of Trump, it’s unclear why they would care.
Dating back to Ronald Reagan’s demonization of “welfare queens,” the GOP has used racially charged rhetoric to undermine support for the social safety net. The result has been over time to empower a demagogue like Donald Trump. A recent New York Times investigation showed that one of the most powerful predictors that a county would vote Trump was its share of citizens living in mobile homes. In the same investigation, the strongest overall predictor of support for Trump was not employment rate, but rather the share of population who were non-Hispanic whites without college degrees. The GOP certainly hasn't done itself any favors by pushing for highly unpopular policies that have benefited their donor class while showing little benefit to the wider population. But, while we accept that all of these factors help explain Trump support, we find that racism is the main driver of support for Trump. The model presented here accounts for all of these attitudes and still finds an incredibly strong relationship between racism and support for Trump. The centrality of racism to the Trump phenomenon should not be obscured.