Skip to content

Can Donald Trump Win?


The short answer is “yes,” but the more qualified answer is “probably not.” Read on for why!

The hot news right now is about the latest Quinnipiac University poll, which shows a tight race between Trump and Hillary Clinton in key swing states. But the poll slants things in favor of white and Republican voters. This is bad news for Trump in a critical way: if the best he can do with polls slanted in his favor is a statistical tie, then he’s got an uphill battle on his hands.

This is good news for Hillary Clinton, of course. She has the advantage. As many have put it, it’s her election to lose. According to RealClearPolitics’ poll averages, Clinton has a healthy 6-point lead over Trump, nationally. But 6% is not a blowout–it’s not a margin that can be taken for granted. It’s breathing room that can easily be lost.

So, let’s start with the possible calamities. What game changers could possibly swing the election easily in Trump’s favor? There are some obvious ones:

  * A terrorist attack. Based solely on the lack of terrorist attacks on US soil in general, I would say this isn't very likely.
  * A new Clinton scandal breaking that manages to gain traction. It should be noted that [Benghazi ]( [emailgate ]( failed to stick. They're popular Republican talking points but don't seem to damage Clinton's electoral prospects more generally.
  * An economic crisis. The economic crisis of 2008 all but guaranteed Obama the election. A crisis of that magnitude in the summer of 2016 would likely see Clinton punished at the ballot box, too. Odds on this aren't terribly worrisome, either, but it's always a possibility.
  * An unusually popular third party candidate. To me, this is the most plausible route to a Clinton loss (and a Trump victory). Anti-establishment fever is in the air, and Clinton is as establishment as they come. Democrat-leaning voters may well decide to cast their lost elsewhere. However, it would probably have to be someone who has a lot of support already--someone like Bernie Sanders, who doesn't seem interested in an independent run.

Any of those events would likely throw the election to Trump. Since they are, to a large degree, unforeseeable at this point, let’s put them aside. Let’s assume none of these things happens, and Trump’s only route to victory is the old-fashioned way: gradually bringing voters around to his point-of-view and voting for him. How does he accomplish that?

538’s Swing-O-Matic ends up being quite a useful tool for this exercise. I’ll just shamelessly steal their description of how it works:

We started with the results of the 2012 election and the support for each party’s candidate by the five demographic groups. We then adjusted the size of those groups based on four years of population change. When you adjust the vote and turnout above, our model recalculates the results for each state — as well as the Electoral College outcome and the national popular vote — taking into account how much of the state’s electorate the group accounts for.

The good news (at least if you’d prefer Trump not win) is that demographics heavily disfavor the GOP, and will continue to do so with each subsequent election due to shifts in the population. They are likely to be denied the Presidency, though their performance at the state and local levels will continue to be robust at least until 2022 (after the next census and redistricting).

So, start with the percentages for each demographic group listed from 2012:

  * College-educated white: 56% Republican, 77% turnout
  * Non-college-educated white: 62% Republican, 57% turnout
  * Black: 93% Democratic, 66% turnout
  * Hispanic/Latino: 71% Democratic, 48% turnout
  * Asian/Other: 67% Democratic, 49% turnout

We can see that there are two dimensions here: vote share, and turnout. Either one can move this year. Taking the 2012 results, that’s 332 electoral votes for Obama, and 206 for Romney. Trump would have to improve on that by a lot: 64 EVs. What shifts does it take for him to get them? At this point I must note that this is all speculation, so take it with a grain of salt. Experiment with the Swing-O-Matic for yourself!

First, consider who Trump’s voters are. I will rely primarily on The Atlantic here, which is using analysis from The Washington Post. Trump’s numbers should improve slightly among college-educated whites, and perhaps a bit more dramatically among less educated whites. I’ll be generous and say a 3% shift in the college-educated white vote (no turnout change). This actually doesn’t move the needle at all in terms of electoral votes–it’s simply not enough. What if non-college-educated white turnout is boosted a few points and leans more heavily Republican? Bumping it to 65% Republican with 60% turnout, you get a narrow Republican victory: Trump gets 272 EVs.

But we’ve yet to look at non-white groups!

I will say that I really have no idea what black turnout will be like. Will some stay home? Will Clinton lose some of those votes, compared to Obama? Some of both seems likely. I definitely don’t see any kind of shift toward Trump. Black turnout was 66% in 2012, and 56% in 2004. Let’s split the difference and say it goes down to 61% but doesn’t shift red. This has a major effect on the electoral votes. Trump now has 315.

It may seem that all is lost, but what about Latino votes? Trump’s rhetoric has been particularly incendiary toward Latino people–referring to Mexicans as “rapists” and “drug dealers.” Statistically speaking, it’s very difficult for a Republican to win without at least 40% of the Latino vote. Polling also indicates that Trump’s support among Latinos is extremely poor–8 in 10 have an unfavorable view of him. So, how much of a swing in the Latino vote would it take for him to hold onto victory, assuming the other shifts I’ve laid out occur?

First, a 3% bump to Latino turnout. It might be more than that, but 3% is about what I’m moving things around so far, so let’s do that. 51% turnout. Electoral votes stay at 315 for Trump. No change. It takes the Latino vote going 77% to the Democrats (with increased turnout) for Trump to see a loss. That’s probably overly optimistic.

Then again, this whole thought experiment hinges on a lot of unlikely events happening. Are working class white voters really going to turn out significantly more and pull for Trump? There’s reason to believe they probably won’t. Worse yet, Trump’s offensive, xenophobic rhetoric probably turns off more educated white voters. Those voters may simply stay home, or even vote for Hillary Clinton. The conservative intelligentsia, such as it is, has moved toward tactical voting for Clinton as a means to thwart Trump.

And if minority voters are actually more mobilized by the prospect of a Trump victory, his chances are totally blown out of the water. In all likelihood, Trump’s victory depends on him mobilizing significantly more voters than would actually support him. Any increases in Republican turnout are probably offset by those who stay home in disgust–or who vote for Clinton instead. For Trump, the only demographic that is likely to boost his vote share is non-college-educated whites, who simply aren’t numerous enough to secure him a win, even if he manages to bring them to the polls.

As I said before, don’t despair. Trump is entirely beatable, because his victory strategy hinges on a demographic that simply can’t deliver him the election. It’s on Clinton not to somehow screw it up, and on the rest of us to vote in November.

Photo by DonkeyHotey