The puzzle of explaining Donald Trump’s support has often been cast in terms of two competing narratives: One of them is about economic anxiety and disillusionment with free trade, and the other is about a racist and/or xenophobic reaction to a changing American demography.
Some liberal commentators, however, have a hard time accepting that both narratives are true, and they're not in contradiction. Last week many seized on a new study to debunk the purported myth that material conditions afflicting working-class white men play a key role in the Trump phenomenon. Trump supporters, Gallup senior economist Jonathan Rothwell found, earn relatively high household incomes. He also found that "living in areas more exposed to trade or immigration does not increase Trump support.”
Vox's Dylan Matthews, in a since-deleted tweet, concluded that "Support for racist demagogue turns out to be primarily driven by racism." Slate’s Osita Nwanevu declared that the study “exposed critical parts of the prevailing narrative about Trump’s rise as myths,” and that his supporters were therefore motivated by “imagined grievances.”
The average “Trump voter,” writes Ryan Lizza at the New Yorker, “is driven not by simple economic self-interest but by something deeper and more psychological.”
But the study concludes no such thing: Trump support, it found, was strongest amongst staunch conservatives, blue-collar workers and veterans' families. Trump voters were likely to live in white enclaves, and in places with high white middle-aged mortality rate and low intergenerational mobility. Rothwell’s study, aside from its much-touted finding about the relationship between Trump support and manufacturing jobs -- which says less than it may seem to at first blush -- therefore lends credence to the argument Matthews and company are eager to reject: Many Trump voters have economic problems that they wrongfully interpret through the lens of xenophobia and racism.
Liberal commentators, overly eager to dismiss the notion that bigotry emerges within class and economic contexts, are ignoring the evidence — such as the fact that Trump and his supporters have been explicit about their economic populism, including their criticism of free trade. In doing so, liberals absolve the Democratic Party’s Clinton-era mainstream for its embrace of immiserating economic policies, and thereby playing a role in Trump’s rise. Democrats should not get off the hook so easily.
Are Trump supporters wealthy?
In his Vox article reflecting on the study, Matthews concludes that “Trump's base is not poor whites — it's way more complicated than that.”
It is more complicated than that. Trump supporters are not better off than other white Republicans, the study finds. They are just higher-income than Americans in general, which is not surprising, since on average white people and Republicans have higher incomes.
Earning a typical blue-collar income does not make one, as Rothwell puts it, “seemingly affluent,” or per Matthews “wealthy” and “doing pretty solid economically.” Wages have stagnated for all working Americans since the late 1970s, and those earned by workers with only a high school education actually fell between 2007 and 2014. While people on the bottom have been devastated, those in the middle aren’t doing so great either and have every reason to worry that their families will be worse off in the future.
I’m not aware of anyone who has argued that Trump’s support is found mainly among American capitalism’s most marginalized people. Rather, the typical argument posits that a combination of economic, nationalist and xenophobic populism is attractive to many white working-class people, men in particular, who face declining economic fortunes and bleak prospects for their children.
What’s more, those people, some of whom work or have worked in factories, are just one piece of the Trump coalition. Some more affluent whites, whose support has received little attention, favor him as well. This, again, is unsurprising: Some rich people have long supported anti-establishment right-wing causes.
Transparent personal economic interests can never fully explain political beliefs. For example, the fact that positive economic indicators are good for the incumbent party holds true not because everyone has an immediate and direct experience of the entire state of the U.S. economy. Rather, regular people’s political beliefs, much like my own or those of Vox writers, are shaped by some combination of their firsthand experience with the economy and their evaluation of how their neighbors and the country as a whole appear to be doing. —The unartful term for the latter phenomenon in political science is “sociotropic voting.” An individual’s economic situation may be relatively secure, but that person may still worry about his or her children and neighbors. Or that person may just be furious about the fact that the share of income taken by the top 1 percent more than doubled, to 20 percent of the total, between 1976 and 2011.
A better question --and an easier one to answer -- is whether Trump supporters say they are motivated by economic concerns. In news articles and opinion polls, they have repeatedly said that they are.
Does economic misery caused by global trade actually have nothing to do with Trump?
“Surprisingly,” Rothwell writes, “there appears to be no link whatsoever between exposure to trade competition and support for nationalist policies in America, as embodied by the Trump campaign.”
Rothwell, however, overstates what his data actually shows.
The study found that people, including white Republicans, who live in areas with a high concentration of manufacturing jobs are actually less likely to support Trump. I asked Rothwell whether this measure might fall into what’s called an ecological fallacy, meaning that it makes inferences about individuals based on measurements of the larger groupings to which they belong. Conclusions drawn about individuals (here, their political preferences) are more likely to be erroneous when the data being used does not directly measure individual-level characteristics but rather the heterogeneous groups or environments to which those individuals belong.
In this case, conclusions about individual-level political preferences are being drawn from data about the heterogeneous areas where manufacturing is concentrated. The nation’s 100 largest metropolitan areas were home to 58.5 percent of manufacturing jobs in 2010, according to a Brookings Metropolitan Policy Program report. Many people who live in metropolitan areas like New York or Portland, Oregon, of course, have never worked in factories.
Rothwell, who engaged me in a detailed discussion of his study by email, says that even if individuals who identify as white Republican manufacturing workers are included in his analysis, it does not predict support for Trump.
This still doesn’t tell us as much as it might seem at first blush. For one, it does not include former manufacturing workers, who ostensibly would be most upset over deindustrialization.
Another important factor is that the manufacturing sector is incredibly heterogeneous, in terms of both its geographic location and its workforce. The fact that certain workers don’t tend to support Trump — say, unionized workers in metro Detroit — does not mean that there are not significant segments of current or laid-off manufacturing workers (and countless neighbors who indirectly depended upon these industries) who do. It’s possible that manufacturing workers in certain areas — large and diverse metropolitan areas with high levels of unionization — are less favorably disposed to Trump, while others feel more favorably.
The “definition of manufacturing is more complicated than surveys make it out to be,” Judith Stein, a historian at City College and the Graduate Center of the City University of New York, wrote by email. “My best example is Hillary Clinton talking about manufacturing by visiting a microbrewery. The attempt to find a single factor for Trump voting, I think, is a fool's errand. Populists have always put together lots of elements in their bills of complaint.”
It is, generally speaking, a very bad time to be a working-class individual. That’s due in part, but not entirely, to role played by global trade in undermining American manufacturing. Workers in other sectors, like mining, have also experienced mass layoffs. And workers across the board have experienced a hit to wages and benefits caused by a sharp decline in union density, by the super-rich hoovering up automation-based productivity gains as profits, and by the decades-long participation of government in corporate America’s coordinated attack on labor rights. These processes affect workers in retail and many lower-skilled service industries just as much as those in manufacturing.
Manufacturing, which represented 19.4 percent of American jobs in 1980 but just 8.5 percent in 2010, is just one piece of the economic fabric that supported living-wage jobs. The bigger point here is that we should not assume the economic bases of Trump support are necessarily about manufacturing per se. No one has argued that most white manufacturing workers, even most white Republican manufacturing workers, support Trump. Rather, the argument is that, as Americans have experienced a jolting shift to a new and precarious economic order, a significant number of middle- and working-class workers have found his attacks on free trade to be compelling.
It’s also worth noting that the decline of unions and the decline in manufacturing are interrelated, and that both have contributed to the rise of the American right. Manufacturing workers are significantly more likely to be unionized than private sector workers as a whole, and union members are much more likely than non-members to support Democrats. And importantly, there are fewer Democrats specifically because fewer people belong to unions, which play a key role in mediating an individual’s political relationship with the state.
There are other geographically specific studies that do find a relationship between trade and the brave new world of American politics — a complicated, multifaceted upheaval in which Trump plays just one part.
According to a working paper by MIT economist David Autor and others, voters in congressional districts exposed to larger increases in import competition were likely to replace moderates with conservative Republicans (in the case of majority-white districts) and liberal Democrats (in majority-nonwhite districts) in 2002 and 2010. Meanwhile, The Wall Street Journal found that Trump won 89 of the 100 counties most impacted by Chinese competition, while Bernie Sanders won 64 of the 100 most exposed Northern and Midwestern counties. (Hillary Clinton, thanks to strong support among black voters, dominated the South.)
What about the racism?
“Trump’s rise has always seemed more closely related to prejudice than economics,” writes Matthews at Vox. “Analysis of surveys has shown consistently that racial resentment correlates more strongly with Trump support than one’s income or degree of pessimism about the economy.”
It’s unclear why Matthews is determined to dub one or the other the main factor. Either way, he misconstrues this data’s meaning.
It’s obvious that racial resentment would correlate most strongly with Trump support: Racist voters can be assumed to be flocking to the most brazenly racist candidate in recent history. By contrast, support among people with low incomes and pessimistic economic outlooks will inevitably be divided among a broader number of candidates. Some economically struggling people are black or Latino. Some are liberal, others are religious conservatives and others are libertarians or Tea Party acolytes. Some might be economically pessimistic but strongly opposed to Trump’s racism. Supporting Trump requires stomaching or embracing both his racism and his brand of economic populism. The working class is not monolithic.
Trump supporters have thrilled to his description of Mexican immigrants as murderers and rapists, his harsh law-and-order language, and his call to bar Muslims from entering the country. It’s obvious that xenophobia and racism are a big part of why people support Trump. But simply declaring “it’s all racism” poses a question rather than answering one: Racism is not a primordial, trans-historical phenomenon but rather an ideology historically embedded within changing social, political and economic contexts.
In the United States, any analysis of racism must begin with black slavery and the genocidal wars of expansion against Native Americans. It continues through imperial wars in Latin America and Asia, and then the Middle East; the rise of Jim Crow in the South and residential hyper-segregation and economic marginalization of nonwhites throughout the country. Finally it must reckon with today’s era of mass incarceration and imperial decline meeting the rise of Islamist terror groups, demographic change and grotesque inequality.
Racism can’t be reduced to an economic function, but it no doubt bolsters certain people’s material privileges while distracting other people from the fact that they don’t have any.
Matthew Yglesias, writing at Vox, concludes that Trump supporters, despite frequently articulating economic grievances, can’t mean what they say because Trump’s economic program is “patently ridiculous” and the “more parsimonious and simple explanation” is that it’s simply “racial resentment,” which he describes as a “basic divide over values and cultural identity.” As if Trump’s insistence that Mexico will pay $5 billion to $10 billion to build a wall to keep Mexicans out is not “patently ridiculous” as well. Given that Trump’s proposals are incoherent and contradictory across the board, it should not be assumed that voters are motivated to support him because they approve of any particular piece of his (shifting) platform.
Trump, an amalgam of far-right populism, is coherently perceived as a giant middle finger pointed at an establishment and status quo that many voters perceive to be a disaster. It’s unclear where Yglesias and company get their folk theory of “culture.” The scholarship on racism contradicts them. To contend that racism is an unfortunate tradition unrelated to this country’s political economy is simply wrong.
It’s really pretty basic: Trump followers like the combination of his racist xenophobia, economic populism and "America First" foreign policy at a moment where profound challenges to U.S. military and economic dominance have thrown the notion of American exceptionalism into crisis.
Why do so many seem so eager to dismiss economic explanations?
I first noticed the inclination to disprove political-economic analysis of Trump in May, when Nate Silver cited a median household income among Trump supporters (who had voted in primaries at that point) of $72,000. This was used to question whether Trump really had working-class support. Though Trump supporters had lower incomes than those backing Ted Cruz or John Kasich, Silver highlighted the fact that Trump incomes were higher than the overall median, and higher than those of Democratic voters.
This is unsurprising: Black and Latino people are the most economically marginalized people in society and tend to vote for Democrats, and Americans in general are less likely to vote the less money they make. And as I wrote at the time: “A two-earner household making $72,000 a year working in construction and dental hygiene is not destitute, can still afford to take a vacation, and maybe even sends their kids to Catholic school. But they’re still working class.”
It is telling that the class-doesn’t-matter narrative is popular among data journalists and commentators who don't do face-to-face reporting, while reporters who interview struggling white people on the ground tend to tell a different story. Holly Otterbein, reporting on her hometown in York County, Pennsylvania, found people who supported Trump because they believed their “grandkids don’t have the same opportunity to build a decent life,” and believe “their lives don’t matter in America — both because they’re working-class and because they’re white.” Trump, she finds, “has also given people someone to blame for their problems,” quoting Michelle Alexander’s observation that the status quo has long been perpetuated by “appealing to the racism and vulnerability of lower-class whites, a group of people who are understandably eager to ensure that they never find themselves trapped at the bottom of the American totem pole.”
Quantitative data is critical to understanding mass social and political phenomena. But it does not speak for itself. Qualitative observations are necessary. And when quantitative data directly contradicts what people observe, it merits close scrutiny. Rothwell’s study is invaluable, and I differ only with his interpretation of manufacturing-related data. The study’s reception from some commentators, however, lacks qualitative evidence and contradicts other quantitative studies.
The debate at hand is not only empirical, however, but political as well.
As Ryan Lizza puts it, “Rothwell’s view is much more in line with the argument that Trump voters are whites who feel that their privileged place in America is threatened by forces they don’t really understand. If this is true, they can’t simply be won over by getting median wages raised or by bringing the local factory back from Mexico.”
The economic argument, Matthews notes, is “a comforting notion, particularly for those on the left. It suggests that large numbers of Americans are not being drawn to a racist demagogue because he’s a racist demagogue, but because of the failures of modern capitalism. It also implies that these voters could be won over by a robust left-wing economic agenda that addressed their plight.”
If there is no economic context, and Trump’s supporters are just mired in primordial racism, then they are forever lost in the morass of right-wing politics. This bolsters the vision of the Democratic Party as comprising an alliance of affluent whites and people of color with a political agenda of multicultural neoliberalism, where economic reforms can be limited to improved educational options and after-tax redistribution. If Trump voters are just “idiots” appealed to “not at a low intellectual level but at a sub-intellectual level” (as Jonathan Chait has put it) then progressives can forget about the angry white guys.
The Bernie Sanders campaign, however, held out another possible future: a multiracial working-class movement with socialist politics, seeking a fundamental reordering of power relations. That Sanders did so well in many white working-class regions where Trump also won big, like West Virginia, made such a strategy seem feasible for the first time in decades, if not longer. This conclusion is bad news for both parties’ establishments and the interests they represent. It’s difficult to believe that it was exclusively the racism or sexism of Democratic primary voters in poor white states that motivated them to support a self-described socialist who likes to cite Denmark as a model country.
We know that Trump finds a lot of support among blue-collar men who live in communities where white people are prematurely dying at shocking rates and young people have a bleak economic outlook. Despite the ways in which he grotesquely builds upon the politics of his more conventional predecessors, Trump is a bizarre and unsettling candidate. We still know too little about what is driving his support. But it’s willful ignorance to insist that economics doesn’t play a major role.
For a long time, leftists have been accused of practicing economic reductionism for contending that working-class people’s conservative political beliefs are shaped by economic conditions they do not understand, and that conservative economic elites have appealed to the proles’ racism precisely to distract them from that material reality. Ironically, large numbers of working-class people and the candidate they support are now articulating xenophobic and racist beliefs to buttress explicitly stated economic pain and anxiety — and some liberal commentators don’t want to hear it. Trump has broken with Republican tradition and, however disingenuously, condemned bipartisan elites for conspiring against the common man’s well-being. It might be comforting for liberal elites to deny this historic reality, but that won’t make it go away.