How Trump Won

Very interesting stuff from Nate Cohn:

In the aftermath of the 2016 presidential election, many analysts suggested that Hillary Clinton lost to Donald J. Trump because of poor Democratic turnout.

Months later, it is clear that the turnout was only modestly better for Mr. Trump than expected.

To the extent Democratic turnout was weak, it was mainly among black voters. Even there, the scale of Democratic weakness has been exaggerated.

Instead, it’s clear that large numbers of white, working-class voters shifted from the Democrats to Mr. Trump. Over all, almost one in four of President Obama’s 2012 white working-class supporters defected from the Democrats in 2016, either supporting Mr. Trump or voting for a third-party candidate.

[…]

Ultimately, black turnout was roughly as we expected it. It looks as if black turnout was weak mostly in comparison with the stronger turnout among white and Hispanic voters.

This was part of a broader national pattern. Mr. Trump’s turnout edge was nonexistent or reversed in states with a large Hispanic population and a small black population, like Arizona. His turnout advantage was largest in states with a large black population and few Hispanic voters, like North Carolina.

What was consistent across most states, however, was higher-than-expected white turnout.

The increase in white turnout was broad, including among young voters, Democrats, Republicans, unaffiliated voters, urban, rural, and the likeliest supporters of Mrs. Clinton and Mr. Trump. The greatest increases were among young and unaffiliated white voters.

For this reason alone, it’s hard to argue that turnout was responsible for the preponderance of Mr. Trump’s gains among white voters. The turnout among young and white Democratic voters was quite strong.

But the turnout was generally stronger among the likeliest white Trump supporters than among the likeliest white Clinton supporters.

[…]

If turnout played only a modest role in Mr. Trump’s victory, then the big driver of his gains was persuasion: He flipped millions of white working-class Obama supporters to his side.

A few points:

• Despite her high personal negatives and her generally negative and character-based advertising, Clinton was able to mobilize the Democratic base about as well as could have been expected. Obama’s only significant advantage was with African-American voters, and it was obviously unrealistic to think that the unusually high African-American turnout in 2008 and 2012 could be replicated, whether Clinton or Sanders or Biden was the nominee.
• The data is not consistent with assertions that Sanders substantially harmed Clinton by endorsing her too late, or too grudgingly, or whatever. Sanders’s core primary constituencies showed up in the numbers that could have been expected or higher and voted for Clinton by the margins that could have been expected. (Stein’s failure to get off the canvas is another indication that Sanders didn’t harm Clinton by not dropping out earlier.)
• The decisive shift of older, higher-income whites without college degrees to Trump is much more plausibly about Trump/Romney than about Clinton/Obama. If Obama had some special appeal to the white working class that Clinton lacked, it certainly wasn’t evident in the 2008 primaries. I could see Biden mitigating some of the defections; Sanders I’m much more dubious although of course nobody knows.
• As I’ve said, assumptions that Trump was a particularly terrible candidate and a generic Republican would have won easily are becoming increasingly problematic. I don’t think it’s safe at all to assume that Rubio or Cruz or Jeb! would have Trump’s particular appeal to white working-class voters.  If the United States had a democratic system for choosing the president, then Trump’s unusual weaknesses would have made him a bad candidate. But in a system that accords undue weight to a few states which had a disproportionate number of voters Trump had a particular appeal to…he wasn’t a weak candidate at all. I’m becoming more and more convinced that a Clinton/non-Trump race would have meant a better popular vote showing but an Electoral College loss for the Republican Party.