Rhetoric like “ Hillary was a complete non-starter for the working class“ and “given the importance of the blue collar vote“ ignores an important point – which is that not all working-class people who are white. And although I don’t have data to offer (none of the exit polls I’ve seen have broken down data by income and race together), it seems very likely that Trump lost badly among non-white working-class voters. (75% of non-white voters without college degrees voted for Hillary). (Source).
The same exit poll indicates that Clinton won among voters with less than $49,000 a year incomes – and her strongest group was those with incomes less than $30,000 a year. Trump won among all economic groups over $50,000 a year. (Possibly a lot of that is because people of color are more highly concentrated among lower income voters).
US median income is about $52,000, btw; so it’s basically accurate to say that Hillary Clinton won a majority of voters with below-average incomes, while Trump won a majority of voters with above-average incomes.
At first glance, Trump’s election seems to have been rooted more in race than in class:
• White voters, who make up 69% of the total, voted 58% for Trump and 37% for Clinton. Non-white voters, who make up 31% of the electorate, voted 74% for Clinton and 21% for Trump.
• White men opted 63% for Trump and 31% for Clinton; white women voted 53% for Trump and 43% for Clinton.
• Among non-college-educated whites, 67% voted for Trump – 72% of men and 62% of women.
• Among college-educated whites, 45% voted for Clinton – 39% of men and 51% of women (the only white demographic represented in the poll where the former secretary of state came out on top). But 54% of male college graduates voted for Trump, as did 45% of female college graduates.
• More 18- to 29-year-old whites voted for Trump (48%) than Clinton (43%).
I think that “working class” and “blue collar” are in effect code; at least when discussing voters, it’s a word people when they mean “white.” Or if that’s not their intention, it’s certainly the effect.
I think the rhetoric about winning classes of voters is as much about the swing as it is about the absolute numbers. The New York Time’s Exit Poll page shows the changes from the 2012 election. For example, Trump got 41% of the under $30,000 income group vs the 25% that Romney did.
Looking at the race/college degree combo category, it seems to mostly have been due to less educated whites shifting towards Trump, with more educated whites shifting towards Clinton. The non-white categories of college/no-college don’t seem to have changed. It’s the shifts, rather than the absolute amounts, that I think indicate the appeal of the individual candidate.
She did massively worse with white people without college degrees, some of whom make less than $50k, so that pulled down her performance relative to Obama in 2012.
We don’t have the stats for it, but I’m pretty sure she did massively worse specifically among white mid-western/Appalachian white people without college degrees (she did ten points worse with women than with men, but that is consistent). She didn’t do massively worse in the South (probably because Democrats already hit the floor for white men without college degrees in the South) or in the rural West, just in Appalachia and the mid-West.
She also did a little worse among Latina women and black men (particularly among younger people) than Obama did, but not enough to make much difference. She also had bad turnout in rust belt cities.
She did appreciably worse among white evangelicals and slightly better among everyone else, relative to Obama. Given that she didn’t do much worse among Southern white evangelicals, I wonder if she did substantially worse among Applachian and midwestern white evangelicals.
She also did badly among voters concerned about immigration and terrorism, but well among voters concerned about the economy and foreign affairs.
In the end, it looks like she is losing within the margin of vote suppression in WI, MI, and PA. MI and PA are CrossCheck states, and WI is it’s own special Hell.
There’s a whole set of cool little graphics about voting demographic behavior over time at the NYT, actually: link
The shift in white evangelicals vs everyone else is interesting. White evangelicals went for Trump by 5% more than they did for Bush (who actually was an evangelical), while everyone else stayed the same in aggregate. I’d be curious to see a map of the shift in Evangelicals and whether the uptick in voting by non-college whites in the mid-west was evangelical (balanced by a drop in evangelicals elsewhere).
Treating “white working-class” as a generic term for The Enemy is part of what brought down the Democratic Party this year, and it’s been holding them back for decades. Maybe their relentless focus on identity politics at the expense of economic issues deserves reconsideration?
“White” is not a dirty word, but you sure couldn’t tell that from reading this year’s Democratic Party Platform.
Food for thought w/r/t the difference between the highly educated and everybody else:
http://whitehotharlots.tumblr.com/post/152963111855/academic-liberalism-has-failed
It’s worth reading the whole thing, but here are some key paragraphs.
As I look at the exit polls that Harlequin linked in #3 what strikes me is that the marriage gap closed considerably this year. Trump won married voters by 10 and lost single voters by 17. However, Romney won married voters by 14 points and lost single voters by 27.
In fact, a great deal of the changes in vote share based on income are probably the direct result of this. The National Election Pool asks about family income, not personal income. Here is a link to the form used in one of the primaries (I can’t find one online for the general, but assume it is still a family income measurement).
The way the income thresholds are set when two people of similar incomes marry they both jump up one (or sometimes two) levels on the Pool’s scale.
This strikes me as the wrong way to look at political trends.
By far the biggest thing about voting is that it is very tribal. The easiest predictor of how anyone will vote is to identify the tribe they voted for last time. For most people that changes at most ONE time in their entire lifetime. Now of course it is a feedback loop–the parties try somewhat to cater to their regular voters. Though they only do so insofar as the party perceives that their votes might have a chance to go to the other side.
When things change, they typically change on the margin of that. MOST people still vote for whomever they voted for last time. But JUST ENOUGH people change that behavior to swing the election.
So when it comes to the working class, it is easy to predict that on a general basis Democrats will have won more of the working class. They did so last election, two elections ago, and three.
Clinton appears to have lost because of two things–getting MANY fewer of traditional Democratic voters to vote for her at all nationwide (including in Democratic stronghold states where voter suppression can’t be a serious enough issue to explain the drop), and showing a very large swing (in the context of the fact that people almost never change sides) against her regarding the working class in the Rust Belt States.
The current meme explaining that among Democrats appears to be ‘racism’. That is a highly problematic claim in general because it doesn’t appear to show the change in voting patterns. (Note that most evidence for this takes the form of “the Republicans harbor racists”. But the actual racists have been voting for Republicans for at least 4-6 presidential cycles so that doesn’t explain the CHANGE).
The Rust Belt States especially show the problem with this explanation: they went for Obama (an actual black man) twice by relatively comfortable margins. For the most part, we fully expect that they will stick with whomever they voted for before–because almost everyone does that. But quite a few of them went from voting for a black man who was running as a Democrat for President twice, to voting for a Republican.
You have to thread the racism needle pretty tight to allow for a big swing of people who were willing to vote for Democrats most of their adult lives, who were willing to vote for a Democratic black candidate for president, and then who suddenly went racist enough to vote for a Republican.
It seems much more intuitive that they changed their minds about something other than race.
Now a much more valid analysis might be that they weren’t put off by Trump’s racism nearly as much as we wish they would have been. They let their personal concerns about job stagnation and job insecurity swamp concerns about Trump’s signaled racism. But that is a VERY different argument from the “they are just racists and I won’t compromise on racism” argument.
Sebastian,
I don’t think we have a good handle on the question of Obama voters voting for Trump, do we? If you have links or detailed analysis, I’d be interested to see it.
Distinguishing between increased turnout for Trump- decreased turnout for Clinton and Obama voters voting for Trump is hard. If someone has done the analysis already to have a strong claim for that, I’m interested in seeing it.
Looking at Michigan exit polls, one point in favor of the idea of vote switchers:
10% of voters thought that the next President should be more liberal than Obama and voted for Trump (i.e. without people who want Trump to be more liberal than Obama, he would have lost by ten if they had just voted Stein and lost by 20 if they’d voted Clinton). Trump also lost liberals by 14% to 81%, while Romney lost liberals 7% to 91%. Trump voting liberals easily swung the vote in MI to Trump.
I just don’t even know how to talk to people like this any more. How the f**k could anyone delude themselves that Trump is more liberal than Obama? How do we deal with people who are that out of touch with reality?