Donald Trump's lead in a number of recent polls -- including one from The Washington Post and ABC News -- is in part thanks to his having at long last consolidated Republican support around his nomination. (He still hasn't mathematically clinched the nomination, mind you, but he keeps getting closer.) But it's also because, so far, he's not doing significantly worse than past Republicans among the non-white voters who are making up an increasing portion of the electorate.
There is a flurry of caveats that are now fluttering down onto that statement, including that 1) there are six months to go, 2) polling of smaller demographic groups means larger margins of error and 3) we rely on exit polls (which are themselves subject to decent-sized margins of error) to make the appropriate comparisons to past voting. Given the role that racial demographics are expected to play this year, though, it's worth seeing how the numbers compare.
Exit poll results reported by the New York Times show how the margins of support for Democrats and Republicans have changed over time for white, black, Hispanic and Asian voters. The figures for 2016 that are shown are from the Post/ABC poll, with the caveat that the sample size for Asian voters was smaller than we would normally report. In each case except among Hispanics, the margin of support is stronger for Trump now than it was for Mitt Romney in 2012. Acknowledging margins of error -- they're all about the same.
Another caveat to drizzle on this: There are still a lot of voters who haven't yet made up their minds, a percentage that for obvious reasons doesn't exist in the exit polling. As a percentage of those who had an opinion, the hollow circles move further away from the center line by a few percentage points for white and black voters, and by slightly more among Hispanics. (You can see that change here.)
That's one survey, but others are within the same range. A Fox News poll released last week showed a slightly higher margin of support for Trump among Hispanics, with Clinton holding a 39-point advantage. Another survey conducted among Asian American voters this spring found that 62 percent of respondents viewed Hillary Clinton favorably -- compared to only 19 percent who viewed Trump in that way.
There is a lot of importance in how much or how little those hollow circles move up or down. As we've noted repeatedly, the percentage of the electorate that is non-white has gotten consistently larger, with Census Bureau data suggesting that the density of the non-white vote in the 2014 midterms was as high as it was in the 2008 presidential race. An uptick in turnout from Hispanics could offset a slight shift back toward the Republicans -- if such a shift were to occur.
Over time, the trend has been that the white vote has gotten smaller and more Republican as the density of votes among Hispanics and blacks has increased and support for Democrats among Asians has dramatically increased. (You can see that on the graph above.) In other words, the history of voting in presidential elections looks like this, using turnout estimates from the U.S. Elections Project and, for 2016, Pew Research. Pew's data suggests that the Hispanic vote will increase, and the white vote will continue to decline.
The lower graph above zooms in on the non-white vote. Which itself serves as a reminder that there's still a wide gap between how many white voters turn out in a presidential general election and how many non-white voters do.
What we're keeping an eye on are the hollow red and yellow circles. Will the red one move up and to the right? Will the yellow one move up and to the left? It's that combination that matters, as this clever tool from FiveThirtyEight makes clear. It allows you to try and figure out how massaging those figures might have changed the election in 2012.
There are white voters who will support Hillary Clinton, of course, and many non-white voters who will back Donald Trump. With five-and-a-half months still to go, predicting how they'll move is tricky -- and, at least for now, they don't seem to have moved much.