Vanity Fair. Horrified that Trump could possibly win, bashing him by noting that he lies "much more" than Hillary (a lie in and of itself) and crediting that bellowing whore as a "statesman" nevertheless hits on some points that polls just don't take into account. This is an older piece, but still makes valid points.
Over the long term, demographic trends favor Democrats, but people vote in the near term. For now, a Mitt-Romney-low level of support among non-whites and a modest increase in the share of white votes would be enough to put Trump in the White House. Freeze the 2012 results and increase the turnout and vote share of only non-college-educated whites by six points, for example, and the results go Republican.
After the election of 2012, we the saw the emergence of a narrative that the Latino vote had been decisive. But this was a myth born of sentiment more than hard numbers. Romney would have required a 40-point gain among Latino voters to swing the election his way, something that was never going to happen. By contrast, a 4-point gain among whites would have had the same effect. Hispanic voters are concentrated heavily in a few states, enough so that California is solid blue, but those states do not include most current swing states.
The battleground states in this election are shaping up to be Florida, Ohio, and possibly Michigan and Pennsylvania. The latter three are states with large white working-class populations and percentages of Latinos at 4.6 percent, 3.3 percent, and 6.6 percent, respectively—all well below the national number of about 17 percent. The exception, Florida, is 23 percent Latino but only 3.5 percent Mexican-American, the group directly insulted by Trump. Twice as many Latinos in Florida are Cuban, and nearly half would vote Republican in a normal year.
Democrats are hopeful that Trump’s rhetoric about Mexico sending its problems to the United States will spur outraged Latinos into the voting booth, and reports suggest this is starting to happen. But even if the spike is large in percentage terms, it’s unlikely to be decisive, because Latino turnout numbers are low (under 48 percent in 2012) and Latino populations are distributed in non-swing states.
Then there’s the black vote. Slate’s Jamelle Bouie has argued that both black and Latino turnout will stay steady or rise in 2016, with Trump doing even worse among African Americans than Romney. Working in Trump’s favor, though, is the fact that Romney did so abysmally among African Americans in 2012—winning only 6 percent of their votes—that Trump has almost nowhere to go but up, and if he goes down, the drop isn’t far. Trump’s focus on immigration and trade is theoretically attractive to a subset of African Americans, many of whom have been negatively affected by both, so there might even be an opening for Trump there. If he makes a serious effort to win their votes and shed his David Duke baggage—pretty much the dumbest suitcase he could have picked up, but Trump is Trump—he might have a chance at winning over an increased African-American vote share, or at least a black voter who isn’t Ben Carson. Again, Trump has little to lose.
This actually goes back to the whole "Republicans better diversify" argument. If you look at the realities presented here, diversification isn't the answer. It's getting the core back.