Can Donald Trump Defeat Hillary Clinton?

It is a question that I’ve heard asked a lot recently, and the intelligentsia has answered “no” — but are they missing something important?

You should beware of all so-called “experts” who claim that a particular outcome is inevitable—who insist one candidate will win and another will lose. The certainty expressed in statements like these (especially those proclaiming Trump has no chance against Hillary Clinton) amounts to little more than bluster and an arrogant attempt to pronounce the future as settled. Naturally, as the “truthsayer,” the speaker imagines they alone know how things must unfold.

Let’s get this straight: The answer to questions like these is always “We don’t know.” When asked, “Can this candidate beat that candidate?” the answer is consistently—and I do mean consistently—“Yes.”

Many people here have declared that Hillary Clinton would not only beat Trump, but beat him handily. That level of certainty is nonsense. Can she win? Would she be the nominal favorite? Is she, at this moment, leading in certain battleground states according to the polls? Yes to all. But just because she can win—and just because she’s polling well now—doesn’t mean she will win, that it’s certain, or that it would be easy.

In fact, there are several realistic scenarios in which Donald Trump could win the election. Before I go further, I should say—since I guess I have to—that I do not like Mr. Trump. I didn’t vote for him in my state’s caucus, and I can’t imagine voting for him under any circumstances. I will never vote for Hillary Clinton—ever—but my distaste for Trump is well-known. Still, it would be foolish—as seems common in many answers here—to let that dislike skew my view of the race.

Let me also emphasize: what follows is merely an illustration of a realistic scenario where Trump could win. I’m not predicting this outcome. This year is too volatile for anyone to predict confidently; so much can change week by week, let alone month by month. Any prediction made eight months out is basically meaningless. If I happened to be correct, it wouldn’t be because I’m clairvoyant—just lucky.

My main point is that a Trump victory is possible. I’ll try to show how that could happen. Below is what I see as a plausible electoral map for a Donald Trump victory.

You’ll notice that I’ve given Mrs. Clinton the benefit of the doubt in states like Nevada, Colorado, Iowa, and New Mexico, which have large Hispanic populations and, in this hypothetical scenario, would support her as they did Obama. There are arguments why that might not happen—such as a surge of blue-collar, middle-class voters—but let’s assume, for this example, that Trump loses those states.

I also granted Mrs. Clinton Virginia, mostly because of the influence of voters in Northern Virginia and how that weighs on the statewide total. Again, there’s a logical counterargument, but just to demonstrate the possibility of Trump winning overall, let’s give Virginia to Clinton.

Even if Clinton wins those states, she still faces a serious challenge in Rust Belt states like Pennsylvania and Michigan, which feature large numbers of white, blue-collar, working-class voters who blame free trade—and countries like China and Mexico—for their economic decline. This also applies to Maine’s 2nd District, which offers only one electoral vote, but could easily go to Trump for the same reasons.

Many of these voters are Democrats who won’t cast a ballot for Clinton—the wife of the president who signed NAFTA—if they’re presented with a genuine alternative. Blue-collar Democrats, especially union members who hate free trade and blame it for declining manufacturing, will gravitate toward Trump. They’ve spent decades lacking a candidate from either party who truly represents their interests. If you’re a millworker who lost your job due to offshoring, you’ve watched both Republicans and Democrats cheer on free trade. You hate it, and you blame it for your problems.

The number of such voters—ignored, dismissed, or ridiculed—is huge, particularly in Pennsylvania, Ohio, Michigan, and Wisconsin. It means those states—especially Pennsylvania and Michigan, which Republicans have tried and failed to flip for nearly 30 years—really could move into Trump’s column, shifting 36 electoral votes away from Clinton.

Meanwhile, the rest of the South and Midwest would likely remain a solid red bloc. Trump might alienate some traditional Republicans, but a reenergized base combined with disaffected blue-collar voters of all stripes could easily counter any loss of the so-called “intellectual right.” Clinton might manage to snag somewhere like Montana, but that’s unlikely, and those three electoral votes wouldn’t matter anyway.

This scenario would give Trump a 290–248 Electoral College victory. Clinton’s only hope, if Trump took Pennsylvania and Michigan, would be winning Florida—which is plausible, given the state’s retirees and diverse population. Still, Florida is fairly red, and it’s not guaranteed. Even if she did win Florida, if Trump took Virginia, he’d still clinch the election. His position becomes that strong when he secures Rust Belt states Republicans usually lose. And that’s without even considering the real possibility he could take a state like Wisconsin, which would add another 20 votes to his total.

In other words, Trump’s path is unique but very real. His appeal to white, working-class voters—who feel ignored and beaten down—is no joke, and Clinton doesn’t have an easy answer for it.

Polling Problems

People who claim Trump can’t win often cite two arguments: (1) Electoral College math—which I’ve addressed—and (2) polling. Let’s talk about polling. Understand what you’re referencing when you say Trump is “losing in the polls” in certain battleground states. Consider how those polls are constructed, whom they’re surveying, and whether the data is actually accurate.

Right now, there’s a crisis in political polling. Polls are becoming less reliable and often predict results incorrectly—even spectacularly so. The core problem is that for a poll to be accurate, it needs to sample the right pool of potential voters. Pollsters rely on assumptions—educated guesses, but guesses nonetheless—about who will actually cast a ballot.

This is especially important now, because response rates to surveys have plummeted over the last decade or two. Pollsters increasingly rely on “modeling,” which requires them to predict what the electorate will look like. If those assumptions are off, the poll is off—sometimes by a lot. For instance, in the 2013 Virginia gubernatorial race, Terry McAuliffe led by double digits in multiple polls just two weeks before Election Day, yet he ultimately won by only 2.5%. Similar polling failures have occurred across the country, in congressional, Senate, gubernatorial, and even presidential races. Overseas, polling also failed spectacularly in the U.K. general election, where the Conservative Party, supposedly headed for defeat, ended up winning a majority.

Pollsters build their models largely on past elections—who showed up to vote, and when. But changing habits, cellphone-only households, entire demographics (often conservatives) refusing to answer polls, and the “Bradley effect” make accurate polling extremely hard. It’s an especially severe problem this year, because the electorate is so volatile. Voters who previously stayed home (or haven’t voted in years) might turn out in large numbers, while traditional Republicans could skip Trump, or blue-collar Democrats might abandon Clinton. The minority voters who came out in droves for Obama may not do so for Clinton. That all adds up to pollsters sampling the wrong people. Whether polls show Clinton leading or Trump leading, they should be taken with skepticism. Recall how Clinton lost Michigan to Sanders, despite being up by almost 20 points in many polls right before that primary. Or how Trump lost Iowa and Maine to Cruz despite polling leads there. Polling, in short, is broken. It should no longer be trusted at face value.

Demographics

Another common claim among Clinton supporters is that demographics guarantee her victory—she’ll win women, Latinos, and African Americans, and those groups will carry her to the Oval Office. While it’s true she may well win each of those demographics, that doesn’t assure victory. Even if she won them by the same margins Obama did in 2012, she could still lose—because turnout is everything. If fewer voters from those groups show up compared to Obama’s elections, and if white, middle-class turnout rises at the same time, the dynamic changes dramatically.

Remember, it’s not just about which demographic shares you win; it’s about overall turnout. Mitt Romney, for example, captured a larger share of the white vote than Ronald Reagan did. But in 2012, white voters were a smaller portion of the total electorate than in 1980 or 1984, and their turnout simply wasn’t as high as in the Reagan years.

Obama drew historically large numbers of young people and minorities to the polls—people who felt a personal connection to his candidacy, some of whom had never voted before. Will Hillary—widely regarded as a weak campaigner, even by her supporters—be able to replicate that enthusiasm? Maybe, but it’s a stretch.

Meanwhile, defecting blue-collar Democrats, especially union members, could be a major threat to Clinton. Do you really think they’ll support a free trader over someone promising to restore American manufacturing by slapping tariffs on imports and protecting domestic industries? If you do, you’re mistaken. That’s a significant liability for Clinton, especially if Trump can unify most Republicans after the convention. It’s not that difficult a task—and if he pulls it off, she’s in trouble.

The Unpredictable Nature of Politics

Finally, politics is inherently unpredictable. You never know what will happen on the campaign trail. Democrats often downplay Hillary’s email problems, but when the FBI is offering immunity to your inner circle over mishandling classified information, you have an issue. Maybe she gets indicted. Maybe Bill has a scandal that rocks the race. Maybe she picks the wrong running mate. Maybe a recession hits in the summer and voters blame President Obama, and by extension all Democrats. A million things could happen. A congressman was once taken down because the FBI found $90,000 in alleged bribe money stashed in his freezer. Politics is weird.

Even the strongest candidates can be derailed by the unexpected—just look at Ed Muskie, Gary Hart, Ted Kennedy. Nixon arguably lost the 1960 election because he sweated too much on television. Never say “never” in a political race.

Could Trump beat Hillary? Absolutely. Could Hillary beat Trump? Of course. It all depends on the circumstances. But this year especially, conventional wisdom, polling, and standard political forecasts are basically useless, because it’s all so unusual. Keep that in mind. Politics is unpredictable, and anything can happen for any reason at any time. Watching campaigns unfold in real time will likely give you a better sense of reality than any set of predictions.