Posted on October 26, 2015

A Whiter Shade of Candidate

Scott Alexander, Slate Star Codex, October 23, 2015

Vox says that Donald Trump practices the “politics of white insecurity”. US News says that Trump shows “the rising power of the white vote”. Salon wants to tell you “eight reasons why white America falls for demagogues like Donald Trump”. The Week says Donald Trump represents “the rise of white identity politics”. The National Journal says Trump is creating problems by “preaching to a shrinking white electorate”.

Read enough of these articles, and you might start to get the feeling that Donald Trump’s supporters are disproportionately white. You would be wrong.

Well, probably. Data are sketchy. There aren’t a lot of polls that sort their questions by race, and when they do there are sufficiently few non-white Republicans that they have trouble getting a good sample size. Nevertheless, the ones we have suggest that Donald Trump’s supporters are about as diverse as any other Republican’s and maybe moreso.

An August YouGov poll with a sample of 30 Hispanic Republicans finds Trump in the lead with that demographic, getting 28% of the vote to runner-up Ben Carson’s 19%. A Gravis poll with about 40 Hispanic Republicans finds Trump with 37% of their vote to runner-up Marco Rubio’s 20%. A Nevada poll also finds Trump leading among Hispanics in that state, though I don’t know their Hispanic Republican sample size. And finally, head-to-head matchups of Trump vs. Clinton show Trump outperforming some past Republican candidates, including Mitt Romney and George Bush, in the share of Hispanic votes he would likely receive.

This picture is confused by articles asserting either that Trump has the highest favorability ratings among Hispanics, or that Trump has the lowest favorability ratings among Hispanics. In fact, both are true! Favorability ratings allow you to rate someone favorable, unfavorable, neutral, or never-heard-of-em. Everybody has heard of Trump, and nobody is neutral about him, allowing both his favorability and his unfavorability to be sky-high (with news sources reporting whichever one of those two facts suits their narrative). Some people have done a little better work and reported his “net favorability”, or favorability-minus-unfavorability, which is very low and indeed negative. But this isn’t what matters in a real election. What matters in a real election is who people vote for. If 40% of Hispanics view him favorably, but they all vote for him, and 60% of Hispanics view him unfavorably, but split their votes among the other ten candidates, Trump has won the Hispanic vote.

As hard as it is to find good data about Hispanics, it’s even harder to investigate black Republicans. The YouGov poll that had thirty Hispanics has only five blacks; it looks like three vote for Carson, one for Rubio, and one for Trump.

Probably more useful are the head-to-head Trump vs. Hillary polls, which survey all blacks (not just Republicans). One finds Trump doing shockingly well and tripling Romney’s (admittedly miniscule, admittedly decreased by opposing Obama) level of support among black voters, but the Washington Post is skeptical and cites others with less extreme results–although even most of those show Trump doing at least as well as Romney and other historical Republicans. In terms of boots on the ground, African-American Daily Beast correspondent Barrett Pitner agrees that “Donald Trump has black supporters–really”, explaining that Trump’s “fear mongering and us-vs.-them tactics have not only created a large supporter base among conservative white Americans, but also black Americans who have been disproportionately hit by the economic downturn.”

There are too few data to say anything for sure. But all of the data that exist suggest that if the Republican primary were held today and restricted to non-whites, Trump would still win. And if Trump were the Republican nominee, he could probably count on equal or greater support from minorities as Romney or McCain before him.

In other words, the media narrative that Trump is doing some kind of special appeal-to-white-voters voodoo is unsupported by any polling data.

On the other hand, there is a candidate whom the media narrative fits like a glove. A candidate who may win primary among whites, but loses in a landslide among minorities. A candidate whose black support is almost an entire order of magnitude lower than his white support.

That candidate is Bernie Sanders.

According to the same YouGov poll mentioned above, 38% of whites support Bernie Sanders for President, compared to 37% of whites who support Hillary for President. However, only 13% of Hispanics support Sanders, compared to 63% for Hillary. And only 4% of blacks support Sanders, compared to 64% for Hillary!

{snip}

Other polls are slightly less extreme but tell the same picture. Gravis (early August) found Hillary leading comfortably among all races, but Sanders’ support among whites was still twice as high as among blacks. The Washington Post also found Sanders doing abysmally overall, but his support among blacks was super-abysmal–only 5 percent!

Suppose we measure a candidate’s “whiteness” by the ratio of their level of white support to their level of nonwhite support within their party. Donald Trump seems to be somewhere around 1.3–1.5. Bernie Sanders is somewhere from 3–10. It isn’t even close. If any candidate is “playing to the politics of white insecurity” or “preaching to the white electorate” or “harnessing the white vote”, it is he.

(though I should clarify that in a general election, Sanders would no doubt garner much higher nonwhite support than Trump just because of the D after his name. We’re only talking about relative to other people in their own party here)

{snip}

But mostly I bring this up not because the presidential primary is interesting in itself, but because it really drives home two important points that I’ve tried to make before.

First, in this post, I suggest that when talking politics “white” sometimes literally means people of European descent, but other times means what I dubbed the “Red Tribe”, very loosely corresponding to Republican voters, but also with connotations of southern, poor, uneducated, religious, and exaggeratedly patriotic. This seems to be one of those second times. Even if Donald Trump had 100% support from all minorities, he would still be “the white people candidate”, or even, as some people have called him, “the white power candidate”. Likewise, even if 100% of Sanders’ supporters were white and no black or Hispanic person had ever had the tiniest positive thought about him, we would never get the same kind of “is Bernie Sanders a demagogue harnessing white voters?” story that Trump inspires every day. Sanders supporters aren’t white! They have degrees from Ivy League colleges! They’re the good guys!

Second, in this post, I argue against the theory that groups with few black members are necessarily racist or exclusive (frequently seen as “Silicon Valley is problematic because of how few black techies there are”). I note that black people are severely underrepresented in groups as diverse as runners, BDSM participants, atheists, fanfiction readers, Unitarian Universalists, furries, and bird watchers. They’re also underrepresented in movements with apparently impeccable leftist and anti-racist credentials, like Occupy Wall Street and the US Communist Party. Given the frequency with which the “your group has few minorities, that means you’re racist and need to become more explicitly leftist in order to shrieve yourself” argument gets used to punch down at nonconformist or “weird” groups, there can never be too many counterexamples. And Bernie Sanders’ campaign is such a counterexample. It fits poorly with the “low nonwhite representation is caused by insufficiently strong social justice orientation” theory, but very well with the counter-theory I propose in that post: nonwhites are just generally less eager to join weird intellectual signaling-laden countercultural movements.

{snip}