Trying to understand the rise of Trump and, in particular, the surge of white working class voters to his banner is a challenge. It is understandable that appalled observers reach for simple explanations that cleanly separate the righteous from those who have strayed into the paths of wickedness. But that doesn't make these simple explanations right, merely understandable as reactions.
David Leonhardt had an excellent column last week that makes this point well.
"Many progressives believe that economic anxiety had little to do with President Trump’s 2016 victory. You can read any number of articles making this case and instead asserting that racism was the reason that a lot of white, working-class voters switched to the Republican side.
I’m skeptical that the story is so simple. I understand why it’s alluring to progressives: Trump is a terrible president, and a race-based explanation for his victory makes his supporters seem less sympathetic than an economics-based explanation. But I don’t think the facts are consistent with the idea that economics were largely irrelevant.
For one thing, the same white working-class voters — the ones who are supposed to be irredeemably bigoted — voted in substantial numbers for Barack Obama in both 2008 and 2012. And even though Trump again stoked racism in the 2018 midterms, many voters flipped back to the Democratic column.
Andrew Cherlin, a sociologist at Johns Hopkins University, has just released a paper laying out an explanation I find more sensible: Racial resentment and economic anxiety both played key roles, and they fed off each other."
Leonhardt goes on to rehearse some of the findings and analysis from Cherlin's paper. Leonhardt does a good job of summarizing but I strongly recommend reading the actual paper in its entirety. It is excellent. Cherlin's paper is organized around a study of two communities, Dundalk, Maryland, a historically white community, and Turner Station, an adjacent, historically black community that, in the past, were heavily dependent on the Sparrows Point Bethlehem Steel plant. He begins the paper as follows:
"The conventional wisdom is that working-class whites supported Trump because of the threat they felt from African Americans and immigrants. Economic factors, it is said, were secondary. But in Dundalk, the long-term decline of well-paying industrial jobs is an important part of the story. To be sure, whites in Dundalk tend to racialize their economic anxieties, but racial issues and the industrial decline are so bound together that it is virtually impossible to separate them into two independent components. The former editor of the local weekly paper, himself a blue-collar worker before becoming a journalist, said:
"If you feel like you’ve got a place in the society around you and your own situation is not tottering on the brink, you’re secure enough to open the door to other people, literally and figuratively. On the other hand, if you’re fearful, desperate, alienated, you start looking for ways to be suspicious of other people."
A former steelworker said, regarding the hostility some people in the community feel towards immigrants:
"When you struggle financially, you don’t have that type of support system. You look for someone to blame. In our world, we’re the hero. We’re not going to be the bad guy, so we’re going to find someone else around to blame it on. If they look like us, that’s harder to accept. I think it’s easy to look out and see somebody that looks different than you and think they’re the problem. Then, when that’s backed up by some politician with an agenda, you feel justified."
Why, then, have several national statistical studies claimed that race was more important than economics? In part because they have under-conceptualized what economics means in these communities: They have largely focused on short-term measures such as current income and employment rather than the long-term decline of the kind of labor that provided people with a sense of pride and dignity. As predictive models (i.e., who supported Trump?), the statistical studies are defensible, but as explanatory models (i.e., what underlies the support for Trump?), they are deficient. Race and economics have been intertwined in these communities for as long as the steel plant operated, and they continue to be intertwined today."
This is a critical point. Prediction is not the same as explanation. This distinction has eluded most of the political science studies on this issue and certainly the many pundits and activists who rely on these studies. Cherlin devotes considerable space in his paper to going over these studies and cataloging their deficiencies in this regard. He accurately describes these studies (I have read all of them) and the assumptions embedded within them.
"The explanation that most observers give for the working class’s embrace of Trump is variously called “identity politics” or “status threat” – a fear of losing ground to blacks and Hispanics – or just plain racism and anti-immigrant fervor. Economic issues, according to this line of reasoning, had little to do with the shift. “After all,” wrote Paul Krugman in the New York Times, “studies of the 2016 election clearly show that racial resentment, not ‘economic anxiety,’ was what put Trump over the top” (Krugman, 2018, Sept. 24). Also writing in the Times, Timothy Egan stated, “In truth, economics will probably not move Trump supporters. Their vote for him was more about status anxiety in a changing nation than about financial uncertainty” (Egan, 2018, Sept. 14). Writing is Vox, German Lopez concluded, “The past year of research has made it very clear: Trump won because of racial resentment” (Lopez, 2017, Dec. 15).
The evidence for these claims is based on academic studies which show that white people’s attitudes toward minority groups were a better predictor of how they voted in 2016 than were their incomes. A number of articles take the form of statistical analyses in which measures of racial attitudes, views of immigrants, or fears of displacement are pitted against measures of economic status, and the former group is found to be more strongly related to voting for Trump than is the latter. The racial measures tap attitudes such as whether respondents believe that whites are discriminated against or whether they think it should be harder for foreigners to immigrate to the U.S. (Sides, Tesler, & Vavreck, 2018). The economic measures tend to be indicators of current or recent financial status such as whether household income has changed or whether the respondent has started looking for work (Mutz, 2018b; Schaffner, MacWilliams, & Nteta, 2018), or whether the respondent reports being in poor financial shape (Cox, Lienesch, & Jones, 2017). Another frequently-cited study concludes that economic distress is not strongly related to the Trump phenomenon because, although his supporters have less education, they have relatively high incomes (Rothwell & Diego-Rosell, 2016).
For Mutz (2018b), the basic reason for the increase in support for Trump is the “dominant group status threat” that many whites feel: The threat to their status as the dominant group has increased because they may lose their position as the majority race in the U.S. population. They also feel threatened, she maintains, by the progress of African Americans, Asians, and Hispanics. Mutz looks at changes in attitudes and in economic circumstances as determinants of change in support for Republicans among a national sample of Americans who were interviewed in 2012 and reinterviewed in 2016. One of her two main measures of status threat is a scale based on responses to questions about social dominance orientation (Ho et al., 2015), such as “An ideal society requires some groups to be on top and others to be on the bottom.” For the other measure of status threat, she uses a scale composed of views on international trade. Opposition to international trade with nations such as China, she argues, reflects, in part, opposition to doing business with countries that are racially different. She finds that changes over time in a person’s attitudes toward social dominance, as well as changes in attitudes toward international trade and China, predicted movement toward support for Republicans, whereas changes in a person’s economic circumstances did not.
Morgan (2018b) reanalyzed the data used in the Mutz article an argued that the results are not strong enough to conclude that the effects of the status threat variables are more important than the effects of the economic variables. He concluded:
"Material interests and her measure of status threat are sufficiently entangled among white voters, especially those in the working class, that it is impossible with her data to estimate their relative importance with any clarity (p. 12).".....
"The problem with trying to separate economic concerns from racially-based identity politics and to determine which is more important runs deeper than debates over statistical modeling. In the studies I have reviewed, and in much of the commentary on racially-based status threat, “economics” is under-conceptualized. Most of the measures used in the empirical analyses have centered on contemporary measures of wages, employment, macroeconomic conditions, and so forth. Without doubt, these are valuable indicators, especially if the aim of the analysis is to predict who supported Republicans or who voted for Trump. Sides et al. (2018), in proposing their “identity crisis” thesis, acknowledge and defend the limitations of their focus on the near term:
"It may seem myopic to focus on short term economic trends, given the longer-term trends toward income inequality. But the impact of inequality on U.S. election outcomes has been ambiguous . . . Less ambiguous, however, is the impact of short-term economic trends, which are strongly related to presidential election outcomes and do help explain oscillating party control” (p. 16).
If the purpose of a study, in other words, is to predict how people will vote in an election, which is the outcome on which the authors are focusing, short-term measures are good predictors. Therefore, as predictive models, statistical analyses of this sort are defensible. As explanatory models, however, they are deficient. There are important long-term [economic] phenomena that can help us to understand why the election outcomes occurred."
Again, explanation vs. prediction. Cherlin concludes his discussion of the white working class population of Dundalk and their movement toward Trump as follows:
"When people in Dundalk are struggling economically, as [newspaper editor] Chuck says, they tend to racialize their anxieties. They talk about economic issues not in the upper-middle-class language of the unemployment rate and the gross national product, but rather in more personal, racial and ethnic terms. They retreat behind their identities and “look for someone to blame,” someone who looks different from them. Racism, the editor argues, is not the driving force. Rather, it’s their desperation about the loss of a way of life, the disappearance of work that gave then pride and dignity, and their sense that no one is paying attention to them. Enter Trump, the politician with an agenda. He listened to the people in places like Dundalk when no one else did. His masterstroke was to recognize the desperation of the white working class over the deteriorating industrial economy and to encourage their tendency to racialize that desperation. Neither economics nor identity politics can be said to be the more important factor. Perhaps one without the other – economics in a setting where no one racialized it, or racial prejudice at a time of economic prosperity – would not have brought about the same result. Together, they were tinder for the bonfire that resulted. And Trump was the match."
Granted this is a more complex story than the simple racism and "status anxiety" story but I believe that it is also more correct and far more likely to lead the left to productive politics than the simple one.
