Among liberals, one of the most popular explanations for Donald Trump’s victory is that it was a “whitelash,” a primal scream of lost influence and latent racism among white voters. I myself certainly talked about racial animus quite a bit during the runup to the election. However, in the spirit of figuring out where we went awry, the actual voting patterns suggest this is flat wrong. Using exit poll data from 2012 and 2016, here is Trump’s share of the vote compared to Romney in 2012:
Whites voted less for Trump than for Romney, while both blacks and Latinos voted more for Trump.1 There’s nothing here that suggests Trump appealed to white backlash in any special way. Quite the opposite. But now let’s add a column to the table:
Among whites, Trump lost 1 percent of white votes, but third-party candidates gained 3 percent. Among Latinos, third-parties gained 4 percent, and among blacks they gained 3 percent.
This is the big difference. Who did third-party candidates hurt the most, Trump or Clinton? And why? Or was the damage equal? You need to answer this question before you can say anything sensible about race.
It’s worth nothing that this doesn’t mean that race played no role in this election. But it does mean two things. First, white racial animus seem to have played no more of a role than it did four years ago. Second, although Trump’s blatant appeal to white ethnocentrism did him little good, it also did him no harm—and that was true among all racial groups. That’s disheartening all on its own.2
When more detailed data is available, it might turn out there are specific subsets of the white vote that moved very strongly toward Trump. But what we have so far doesn’t suggest anything of the sort. If you still want to claim that whitelash played a big role in this election, you need to contend with this.
1You can break this down by age or gender, but it doesn’t really change anything. For example, white men moved slightly toward Trump while white women moved slightly away from him. Likewise, middle-aged whites moved slightly toward Trump while young and old whites moved slightly away. But the differences are small enough that they don’t change the picture much.
2Since I first put up this post, several people have suggested that national data isn’t the right way to look at voter demographics. Instead, we should look at the key swing states of Pennsylvania, Wisconsin, and Michigan. But that doesn’t change things. If you look at the exit poll data, Trump did slightly worse than Romney in Pennsylvania and slightly better in Wisconsin and Michigan. But the operative word is “slightly.”
Still, maybe turnout was up among white voters? That’s possible. But we don’t have that information yet, and I’m not sure when we’ll get it.