The Washington PostDemocracy Dies in Darkness

A massive new study debunks a widespread theory for Donald Trump’s success

(Rachel Orr/The Washington Post)

Economic distress and anxiety across working-class white America have become a widely discussed explanation for the success of Donald Trump. It seems to make sense. Trump's most fervent supporters tend to be white men without college degrees. This same group has suffered economically in our increasingly globalized world, as machines have replaced workers in factories and labor has shifted overseas. Trump has promised to curtail trade and other perceived threats to American workers, including immigrants.

Yet a major new analysis from Gallup, based on 87,000 interviews the polling company conducted over the past year, suggests this narrative is not complete. While there does seem to be a relationship between economic anxiety and Trump's appeal, the straightforward connection that many observers have assumed does not appear in the data.

According to this new analysis, those who view Trump favorably have not been disproportionately affected by foreign trade or immigration, compared with people with unfavorable views of the Republican presidential nominee. The results suggest that his supporters, on average, do not have lower incomes than other Americans, nor are they more likely to be unemployed.

Yet while Trump's supporters might be comparatively well off themselves, they come from places where their neighbors endure other forms of hardship. In their communities, white residents are dying younger, and it is harder for young people who grow up poor to get ahead.

At a rally in Detroit, Republican presidential candidate Donald Trump outlines what he would do as president to take the U.S. economy to "amazing new heights" (Video: Sarah Parnass/The Washington Post)

The Gallup analysis is the most comprehensive statistical profile of Trump's supporters so far. Jonathan Rothwell, the economist at Gallup who conducted the analysis, sorted the respondents by their Zip code and then compared those findings with a host of other data from a variety of sources. After statistically controlling factors such as education, age and gender, Rothwell was able to determine which traits distinguished those who favored Trump from those who did not, even among people who appeared to be similar in other respects.

Rothwell conducted this kind of analysis not only among the broad group of Americans polled by Gallup. He was also able to focus specifically on white respondents, and even just on white Republicans. In general, his results were the same regardless of the group analyzed.

Rothwell's research includes far more data than past statistical studies of Trump. It also provides a detailed view not only of the people who support him but also of the places where they live. Academics and other analysts will continue to study the Trump phenomenon in months and years to come, and may, of course, reach different explanations.

This research leaves some mysteries unsolved. Something is afflicting the places where Trump's supporters live, but Trump's supporters do not exhibit more severe economic distress than do those who view him unfavorably. Perhaps, Rothwell suggests, Trump's supporters are concerned less about themselves than about how the community's children are faring. Whatever it is, competition from migrant labor or the decline of factory work appear to be inadequate explanations.

Trump is giving his supporters a misleading account of their ills, Rothwell said. "He says they are suffering because of globalization," Rothwell said. "He says they’re suffering because of immigration and a diversifying country, but I can’t find any evidence of that."

Trump's support does come from a place of adversity, though, and Rothwell said Trump's prescriptions — tariffs on imported goods, restrictions on immigration and mass deportation — seem disconnected from his voters' real problems.

"I don’t see how any of those things would help with their health problems, with the lack of intergenerational mobility," Rothwell said.

Five findings in particular from Rothwell's work are noteworthy: those related to economic factors such as income, manufacturing and opportunity, as well as his conclusions about health and racial diversity.


From polls, it is clear that Trump's supporters tend to be blue-collar men with lower levels of education. Yet important questions remain. For instance, do these people support Trump because they are on the margins of the economy or for other reasons?

To answer these questions, Rothwell gathered data, mostly from Gallup's regular telephone interviews. In those interviews, pollsters asked how favorably respondents viewed the presidential candidates and collected a variety of other information, including where respondents lived, their race and ethnicity, their religion, their education, their employment and their income. Rothwell also compiled information about the communities where people lived — how healthy the residents were, the local effects of trade, and the level of economic opportunity. He compared all these factors to determine which were closely associated with Trump's supporters.

Among people who had similar educations, lived in similar places, belonged to the same religion and so on, those with greater incomes were modestly more likely to favor Trump. They were just as likely to be either working or looking for work as others.

In one respect, that conclusion was expected. White households tend be more affluent than other households, and Trump's supporters are overwhelmingly white. The same is true of Republicans in general. Yet when Rothwell focused only on white Republicans, he also found that demographically similar respondents who were more affluent viewed Trump more favorably.

These results suggest that personal finances cannot alone account for Trump's appeal. His popularity with less-educated men is probably due to some other trait that these supporters share.


Several recent analyses have attributed Trump's success to the disappearance of the factory worker, and to competition with imported goods — especially from China. An essay in the Atlantic in May attributed Trump's success to the gradual decline of employment in the manufacturing sector because of technology and globalization.

"Manufacturing provided steady work for unionized workers without a four-year diploma," Derek Thompson wrote. "When it collapsed, so did unions and the fortunes of non-college men."

On Thursday, a Wall Street Journal report was published online with the headline "How the China Shock, Deep and Swift, Spurred the Rise of Trump." The authors concluded that Trump had won the Republican primary in 89 of the 100 counties most negatively affected by competition from China, measured according to an index developed by a group of academic economists.

Trump's supporters do live and work in economies reliant on manufacturing that have been exposed to intense competition from China. They themselves believe their personal finances have been negatively affected by trade: A poll by the Pew Research Center during the primary found that 60 percent of Trump's supporters said trade had hurt their family's finances, compared with 42 percent of Ohio Gov. John Kasich's supporters and 36 percent of those supporting Sen. Ted Cruz (R-Tex.).

Yet the Gallup analysis shows that Americans who live in places where employment in manufacturing has declined since 1990 are not more favorable to Trump. Rothwell did not find a relationship when he focused only on white respondents, either, or even specifically on white Republicans.

Trump's supporters have many other traits in common with the factory workers whose economic prospects have been negatively affected by automation and global trade. They tend to be less educated men who hold blue-collar occupations.

Yet those two broad trends in factory work do not account for Trump's appeal, Rothwell's analysis suggests. In fact, among those who share other traits, those who live in districts with more manufacturing are less favorably disposed toward Trump.

Rothwell even found that evidence that people in places affected by Chinese competition viewed Trump more unfavorably. Rothwell, however, was less confident in this finding because of statistical uncertainty.

Rothwell said the results make sense, even though he was surprised by them initially.

Trump's supporters are blue-collar, and many people working in those occupations have jobs in construction, repair or transportation — all of which are protected from Chinese competition. Chinese workers might be assembling semiconductors, but they are not adjusting the thermostat or changing the oil.

Republicans who belong to unions outside of the public sector are not more likely to favor Trump than those who are not in a union, but self-employed Republicans view the candidate more favorably, after adjusting for other factors.


Trump supporters might not be experiencing acute economic distress, but they are living in places that lack economic opportunity for the next generation.

Rothwell used data from Harvard economists Raj Chetty and Nathaniel Hendren, who studied how children born in the 1980s moved up or down the economic ladder depending on where they grew up. Children raised in places with high economic mobility, such as Boston or Pittsburgh, often surpassed their parents in socioeconomic status. Children raised in places with low economic mobility, such as Raleigh, N.C., and Indianapolis, struggled just to do as well as their parents in adulthood.

Trump was especially popular in these parts of the country.

Why does Trump’s message resonate the most in these low-mobility areas? The data do not provide a clear answer. It is possible that Trump's supporters, while still better off than many of their neighbors, are worse off than they might have been in the past. Rothwell examined their incomes, but he did not have data on how those incomes had changed over time.

Polling conducted by The Washington Post and ABC News earlier this year, for example, also found no connection between current income and support for Trump. Respondents were also asked, however, whether they felt they were struggling to maintain their standard of living or whether they felt comfortable in their situation and that they were moving up. Those who said they felt they were struggling were more likely to support Trump.

Economic and racial anxiety: Two separate forces driving support for Donald Trump

Rothwell also suggested the reason might have something to do with parents and children. Trump voters tend to be older, blue-collar workers, and recent generations have had more difficulty getting well-paying jobs that didn’t require much education. Those opportunities have largely dried up. And now, Trump supporters tend to live in places where the world has gotten visibly tougher for the kids on the block. It's easier to agree with Trump's narrative about American decline when you have seen your own child fall down the economic ladder.

This may help explain one puzzle that has stumped election observers so far. Trump has found success playing up economic grievances, stoking anxieties about immigrants, and complaining about Chinese competition. How is it then, that so many of his supporters seem to be economically secure? It could be that Trump supporters aren't worried for themselves, but for their children.


As The Post reported in March, the counties that supported Trump in the GOP primaries were the same counties in which middle-aged whites suffer from abnormally high death rates. Rothwell's report confirmed this connection and expanded on it.

Death predicts whether people vote for Donald Trump

Among Americans who were similar in terms of income, age, education and other factors, those who lived in places where people were less healthy had more favorable views of Trump. In these communities, whites are dying faster, there is more obesity, and people report more health problems. Again, this pattern held when Rothwell focused on white respondents only and on white Republicans specifically.

In other words, between two people who earn the same amount of money and have the same amount of schooling, the person who comes from a place with bad health is more likely to support Trump. It's hard to say what is causing this bad health, but at least some of this probably has roots in cultural practices — diet and exercise habits, patterns of drinking and smoking, and more.

It’s unclear what’s going on here, but it’s not a recent phenomenon. Economists Anne Case and Angus Deaton recently documented startling increases in the middle-aged white death rate in the past decade, but Rothwell finds that people's support for Trump didn't seem to be affected by changes in the white death rate where they lived. The places where Trump is popular are places where people have been unhealthy for a long time.


Although Trump voters tend to be the most skeptical about immigration, they are also the least likely to actually encounter an immigrant in their neighborhood.

Rothwell finds that people who live in places with many Hispanic residents or places close to the Mexican border, tend not to favor Trump — relative to otherwise similar Americans and to otherwise similar white Republicans.

Among those who are similar in terms of income, education and other factors, those who view Trump favorably are more likely to be found in white enclaves — racially isolated Zip codes where the amount of diversity is lower than in surrounding areas.

These places have not been affected much by immigration, and Rothwell believes that is no coincidence. He argues that when people have more personal experience of people from other countries, they develop friendlier attitudes toward immigrants.

Research from Pew suggests that there is a relationship between the character of people’s neighborhoods and their views on immigrants. A study from 2006 found that native-born Americans living in Zip codes with lots of immigrants tended to hold immigrants in higher esteem. For instance, they were about twice as likely to say that immigrants “strengthen the US with their hard work and talents.”

This was true, apparently, even after taking into account people’s backgrounds and their political leanings. “Analysis of the survey indicates that their more favorable views do not merely reflect their demographics or political composition, but suggests that exposure to and experience with immigrants results in a better impression of them,” Pew noted.