Skip to main content
ABC News
How Trump Could Win The White House While Losing The Popular Vote

OK, before I say anything, a quick disclaimer: This piece is not a prediction. In fact, I’m a religious (maybe fanatical) adherent of FiveThirtyEight’s 2016 election forecast model, which I find to be both methodologically rigorous and intellectually honest. I don’t dispute its assessment that Hillary Clinton has a 63 or 64 percent chance of winning the election.

That said, in the event this race does tighten to a coin flip by Nov. 8, there is an unusually high chance Donald Trump could win the Electoral College while losing the popular vote — basically, Democrats’ version of the apocalypse.

Here’s why: Several of Trump’s worst demographic groups happen to be concentrated in states, such as California, New York, Texas and Utah, that are either not competitive or that aren’t on Trump’s must-win list. Conversely, whites without a college degree — one of Trump’s strongest groups — represent a huge bloc in three blue states he would need to turn red to have the best chance of winning 270 electoral votes: Florida, Ohio and Pennsylvania.

A repeat of 2000’s split verdict — except with more potential to plunge this much more polarized and anxious country into chaos — is still not very likely. Right now, the FiveThirtyEight polls-only model posits a 6.1 percent chance of Trump winning the Electoral College while losing the popular, and a 1.5 chance of the reverse outcome. But that’s not so remote, either, and if the national ballot were ever to tighten further, both “crazy” scenarios’ odds could rise.

The secret to how Clinton could win more votes nationally yet still fall short of the White House lies with Trump’s weakness among three geographically disadvantaged groups of voters:

1. College-educated whites. Poll after poll shows Trump performing abysmally for a Republican among whites with a college degree, particularly women. But according to my analysis of census and exit poll data, whites with a degree exceed 35 percent of the likely electorate1 in just a dozen states (plus Washington, D.C.): Colorado, Connecticut, Kansas, Maine, Massachusetts, Minnesota, Nebraska, New Hampshire, New York, Oregon, Rhode Island and Vermont.

Of those states, only Colorado and New Hampshire are considered swing states, and Colorado doesn’t even look that competitive right now. Trump would love to win New Hampshire, but he doesn’t need to carry it in order to obtain 270 electoral votes — it ranks relatively low on FiveThirtyEight’s tipping-point ratings, with only a 2.9 percent chance of providing the winner the decisive electoral vote. So if Clinton wins historic margins among well-educated whites, they could help her win by millions of additional votes in states that won’t decide the Electoral College victor.


2. Hispanics/Latinos. According to the Census, only 48 percent of 23.3 million eligible Latinos turned out to vote in 2012, and the Pew Research Center estimates 27.3 million will be eligible this year. With Trump atop the ballot, Latinos could be poised to break records for turnout and Democratic support. But they’re woefully underrepresented in Electoral College battlegrounds and provide Clinton with millions of superfluous votes in California, New York and Texas.

Of the battlegrounds, Latinos exceed 15 percent of the likely electorate in only Arizona, Florida and Nevada. Trump doesn’t need Nevada, and by itself, a Latino surge likely won’t be big enough to erase the GOP’s 9 percentage point 2012 margin in Arizona. By far, Latinos are most potent in Florida. However, Florida was Obama’s narrowest win in 2012, and even if Trump were to underperform Mitt Romney by 5 points among Florida’s Latinos, he could flip the state by winning whites by an additional 3 points.


3. Mormons. According to the Pew Research Center, Romney won fellow Mormons 78 percent to 21 percent in 2012. The LDS church counts 6.5 million U.S. members, and if Mormons voted their weight, Romney probably carried them by about 1.5 million votes. However, Trump is massively unpopular among Mormons, and it’s entirely possible he could win them by just 10 points over Clinton, with many opting for Libertarian Gary Johnson or independent Evan McMullin instead.

If that many Mormons defect from the GOP, it could effectively shift the national popular vote by 1.3 million in Democrats’ favor — more than twice Al Gore’s margin in 2000. Yet the exodus seems unlikely to net Clinton additional electoral votes. Trump is on track to win a plurality in Utah, and outside the Beehive State, the bulk of Mormon voters are concentrated in noncompetitive Idaho and California. They may only matter on the margins in Arizona and Nevada.


Using a prototype of a demographic election calculator that FiveThirtyEight will be unveiling in the next few weeks, I decided to simulate a few election scenarios. Starting with 2012 results as a baseline and adjusting for demographic changes over the past four years, I tested what the map would look like if African-American turnout dipped, GOP support among college-educated whites and Latinos slightly declined, and noncollege whites rallied to Trump in large numbers.

More specifically, here are the conditions I used to set up a fairly plausible scenario that would scare the heck out of Democrats:

  1. Latino turnout rises from 48 percent in 2012 to 54 percent, and their support for Democrats increases from 71 percent to 74 percent.
  2. Asian/other turnout rises from 49 percent in 2012 to 54 percent, and their support for Democrats increases from 69 percent to 74 percent.
  3. African-Americans continue to give Democrats 93 percent of the vote, but their turnout falls from 66 percent to 60 percent.
  4. Among college-educated whites, turnout remains steady at 78 percent and Republicans’ share falls from 56 percent to 47 percent.
  5. Among whites without a college degree, turnout surges from 55 percent to 66 percent and Republicans’ share rises from 62 percent to 67 percent.

The result? Clinton would carry the popular vote by 1.5 percentage points. However, Trump would win the Electoral College with 280 votes by holding all 24 Romney states and flipping Florida, Ohio, Pennsylvania, Iowa and Maine’s 2nd Congressional District from blue to red. And the real disparity between the electoral and popular votes could be larger, because this model doesn’t even factor in Trump’s Mormon problem.

Don’t get me wrong: This scenario is still very unlikely. But its potential to plunge an already fraught election into absolute chaos means it shouldn’t be discounted, either.

FiveThirtyEight: Election Forecast Update – Sept. 15, 2016


  1. Thirty-five percent isn’t some magic cutoff, but it works to highlight the states where college-educated whites are a disproportionate electoral force.

David Wasserman is the U.S. House editor for the Cook Political Report.