How Trump Wins

A Trump surprise is not impossible. Let’s start with a premise: on the morning of November 9, we wake up to an America where Donald Trump has been elected President of the United States. Let’s also assume there was no catastrophic moment for the Clinton Campaign between now and then. Considering electoral projections on October 27, how did that happen?

An important piece of this puzzle would be a small Trump rally in the polls, which is feasible. Only 80 percent of Republicans in yesterday’s national Fox News poll backed him, leaving a fifth of GOP voters to come home between now and November 8. If we see that gradually happen over the next 12 days, Trump can claim some momentum and potentially convert that into wooing undecideds. The media will push the comeback narrative since a closer horse-race is good for ratings, and it will also look for ways to jettison its reputation as an arm of the Clinton Campaign.

While this reversal of momentum will help, it won’t be enough. The national polling gap is too large. Or is it? Whether looking at four-way or two-way polls, Trump might be closer than recent stories suggest. Hillary Clinton’s advantage, which several times reached double digits after TapeGate, has dropped to about 5 or 6. Two recent polls have Trump down just one point, a statistical tie. One such poll is the IDB/TIPP, which, with Pew, boasts the most accurate presidential results of the last four election cycles. In the four-way race, it has Clinton 44, Trump 43 (Johnson 4, Stein 1).

We might also learn that we should have had a lot more faith in the LA Times/USC Tracking poll, which tracks the same group of voters across the election rather than relying on different samples. Voters were chosen based on who they voted for in the last election, and they give their responses exclusively online. While some argue against that kind of self-identification (people often lie and say they voted for the winner) and online polling (preferring live phone calls), there are arguments to be made in favor of it. Gauging the same voters across the election rather than sets of people who decided to answer their phone that day might offer a better indication of how the races is progressing. Online polls also offer respondents the same anonymity that a voting booth does; in other words, anyone sheepish about their Trump vote might not concede that to an actual person, but they might admit it to a checkbox on a computer. This type of tracking poll, it’s worth noting, almost nailed the 2012 results. Although others showed a near tie in the race, this methodology predicted a 3.32 percent win for President Obama, pretty close to his 3.86 margin in the actual results.

As of yesterday, the LA Times/USC poll had Trump up by one point.

There are other factors to consider. Undecideds and third party support are hanging around, though they predictably dwindled. Real Clear Politics’s four-way average now has the two major candidates combine for about 85 percent of the vote. The other 15 percent of the country is split between Jill Stein (2 percent), Gary Johnson (6 percent), and Undecided (7 percent).

How much of that 15 percent will dwindle further? Consider that in the 2012 election, Obama and Romney combined for 98.3 percent of the national vote, and in 2008 Obama and McCain pulled down a healthy 98.6. Though Real Clear Politics polls had both candidates totaling in the low-to-mid-90s on the eve of the election, there was, in both cases, a spike at the end, culminating in the ultimate poll — the election itself. If recent history repeats, we can expect this year’s 15 percent for “other” to fall as well.

Let’s say it falls to 10, which would still be double the highest recorded “other” since Ross Perot. The question then is — to whom do those voters go? If we’re assuming a Trump victory, we have our answer. As for why they went to him, the answer is probably that third party backers are clearly antiestablishment and antiClinton, and voting for Trump gave them the opportunity to promote that. We would also find that Undecideds, who know Clinton well and haven’t supported her for that reason, were just looking for a reason to vote Trump and he gave them one down the stretch. Supporting the potential for this development is that the last two national polls have Trump winning independents, 41-28 and 40-35.

Finally, early voting in crucial swing states suggest Trump is more competitive than national polling suggests. In the biggest battleground state of Florida, returned Republican ballots outnumber Democratic ones. In ever-important Ohio, requests for early ballots from black voters are down 10 percent, while requests from white voters are up 3. In a state that’s perpetually close, that could make the difference.

In total, the above factors point to an unmeasured movement. His enormous rallies are suggestive of the same, though they’re not indicative of one and no one would call them the silent majority (or silent anything else). However, if:

• the most accurate poll in recent history says Trump is down only one;
• another accurate, potentially more reliable poll says Trump is up one;
• fifteen percent of the country hasn’t chosen between the two yet, which gives us a sizable variable;
• independents are breaking toward Trump;
• and the earliest actual votes imply Trump can take home two enormous battleground prizes;

then we might indeed wake up on November 9 with a President-elect Trump.

That’s how Trump wins.