I will admit that the following paragraph from Matthew Dowd writing at the WSJ is not well constructed, so perhaps I am missing something, but he seems to have an odd definition of low demographic diversity (at least in terms of a likely Clinton electoral coalition):
2. On Election Day it is highly likely that the winning coalition of each major-party nominee would set a record low in demographic diversity. Mr. Trump could win this election with 43% of the total vote (due to a large percentage of votes going to third- and fourth-party candidates) by netting fewer votes among non-whites than any candidate has in a century. It is possible for him to win this election with only 20% support among all non-whites. Mrs. Clinton could win by netting fewer white votes than any candidate since Walter Mondale or George McGovern lost in landslides. She could win even if she gets only a third of the white vote. The larger issue is, as with the candidates’ low trust and likeability numbers, that these data points could figure in an election victory but do not set up a standard for governing a diverse country from coast to coast. Some large group of voters is going to feel left out no matter who wins.
Now, it is true that a Trump win would very likely be dominated by white males (and would overwhelmingly be white). As such, if that were the outcome, it would indeed reflect a winner with very low demographic diversity. However, if Clinton wins with only a third of the white vote (as per Dowd’s hypothetical) then that would mean that she won with a mix of white, black, Hispanic, and other voters, making it perhaps the most diverse winning coalition of all time. Yes, if Clinton wins a lot of white male voters will feel left out, but that is one group feeling left out, while a Trump win would mean white women and blacks, Hispanics, and others of all genders would feel left out.
Maybe the error is in a sloppy headline to the paragraph, but it comes across as defining “demographic diversity” as requiring a huge chunk of white males, which is an odd definition.
Still, as this poll from May helps illustrate, Clinton does well with an array of voters, while Trump only dominates a few categories: Race, Gender Biggest Differentiators in Views of Clinton, Trump.
This is also illustrated in a piece from Nate Cohn from late July: The One Demographic That Is Hurting Hillary Clinton:
The list of voting groups generally alienated by Donald J. Trump is long: Hispanics, women, the young, the college educated and more. How is it that he’s in such a close race with Hillary Clinton?
The answer lies with a group that still represented nearly half of all voters in 2012: white voters without a college degree, and particularly white men without a degree.
Mrs. Clinton is showing enormous weakness with this group. And these voters are supporting Mr. Trump in larger numbers than they supported Mitt Romney four years ago. It’s enough to keep the election close. It could even be enough for him to win.
Of course, apart from raising questions about Dowd’s use of terms, all of this continues to underscore how a major issue for American politics going forward is going to be addressing the way in which demographic shifts are affecting political behavior (as well as creating a myriad of challenged for both parties).
