As we chug toward 2020, this is a lesson that’s important to remember — particularly as we observe now-President Trump regularly asserting that he’s a lock for another four years. Or, he “jokes,” many more years after that.
On Friday morning he became incensed at a column by the New York Times’s Tom Friedman, which Friedman somehow managed to type while vigorously wringing his hands.
“Trump’s Going to Get Re-elected, Isn’t He?” the column, published earlier this week, is titled. It’s a stew of what you’d expect from a Times columnist: These are the Democrats who are challenging him? But what about the economy? There’s a perpetual fatalism that hangs over Democrats like a mopey cloud, and Friedman sees only rain.
Trump, true to form, decided to attack Friedman on Twitter since the article calls him racist. It’s a weird tactic, since the number of Trump voters — or voters overall — who read Tom Friedman columns is certainly fairly small. (Trump’s choice to disparage Friedman for playing a lot of golf has spurred Alanis Morissette to return to the recording studio for an update.) Trump should celebrate this idea that even centrists think he’s going to win again, but that’s not how Trump works.
Let us then offer a counterpoint to Friedman’s assertions, one that Trump should be frustrated by. What are the indicators which suggest that perhaps Trump won’t easily win reelection?
Let's start where Friedman ends.
Is the economy a guarantee?
At Axios on Friday, a data point was offered that suggests that Trump is well-positioned on the economy: “Every incumbent president since FDR has won if he avoided a recession in the lead-up to an election year.”
Well, that's that, right? Can't argue with data.
Except that the sample size here is . . . small. There have been 13 presidents since Franklin Roosevelt. Only three have lost reelection bids: Gerald Ford, Jimmy Carter and George H.W. Bush. There have been seven recessions since 1970, two of which ended in the first quarters of the year before reelection bids (Ford’s and Bush’s). Is that what doomed them? More important, does that mean that a president without a recession can’t be defeated?
There’s a great XKCD comic that gets at the problem with extrapolating from past results. There are so few presidents and so many factors involved in the national economy that, regardless of the moment, a seemingly impenetrable pattern can be put together that no president (or no modern president) has broken. Until it gets broken.
Consider another of those patterns: No president with approval below 50 percent has been reelected. Unlike Ford, Carter and Bush, Trump has never been above 50 percent in Gallup polling. Is he doomed?
We have a political experiment that can help answer this question: the 2018 midterm elections. The pattern based on the economy suggested that Republicans should have lost a handful of seats last year. The pattern based on Trump’s approval suggested that his party should lose about 40.
You probably remember which one happened.
Will Trump win because of turnout?
It’s obviously true that Trump is president because more of his supporters turned out in enough places in 2016 to give him an electoral vote victory. It’s also true — though Trump tries to deny or deflect from it — that he lost the popular vote by a wide margin. He is president, thanks to fewer than 78,000 votes in three Midwestern states but was America’s second choice overall by nearly 3 million votes.
At the New York Times, Nate Cohn has an interesting analysis of how a surge in turnout might affect 2020. Historically, a boost in turnout has been a boon to Democrats since voters who vote more regularly tend to be older and wealthier — and therefore more Republican. Next year may be different, Cohn argues, because Trump is popular with Republican-leaning voters who also don’t turn out heavily unless they feel inspired to go to the polls for Trump — a push they may not have felt in 2018.
That’s fair. But it is also worth highlighting the other side of that equation: those infrequent Democratic voters.
A study completed last year looking at voter files determined that about 4.4 million people who voted for Barack Obama in 2012 didn’t vote at all in 2016. Slightly more than that number voted for Obama in 2012 but Trump in 2016. But while Obama-to-Trump voters were mostly white, the majority of the 4.4 million non-voters weren’t.
Why does that matter? Because nonwhite voters are much more loyal to Democrats.
Geography matters in presidential races, as 2016 demonstrates. Those 4.4 million voters weren't necessarily in the three states where Trump won a narrow majority, but it's hard to imagine that turning out another 2.2 million nonwhite voters wouldn't have potentially shifted the result in, say, Michigan, which Trump won by about 11,000 votes.
Or, more interesting: What about Texas? It’s likely going to support Trump in 2020, but, of all of the states he won in 2016, it’s also the one where his approval rating last year was the lowest. A surge in turnout among nonwhite voters or those who stayed home in 2016 could be significant — particularly if coupled with broader apathy from infrequent Trump voters.
It’s hackneyed to say that it all comes down to turnout. But in 2016, it did. We’re far enough out now that the turnout picture is extremely muddy, but it’s safe to assume that it’s not guaranteed to aid Trump’s reelection.
Will America give him the benefit of the doubt again?
There’s another factor that was important in 2016. Trump was a deeply unpopular politician running against another slightly less-deeply unpopular politician, Hillary Clinton. People who disliked Trump but liked Clinton voted for her. People who disliked Clinton but liked Trump voted for him. People who liked both of them? Well, there weren’t enough of those voters to tell us anything.
There were, however, plenty of people who disliked both of them. Nationally, Trump had a 17-point edge with those voters, according to exit polls. In the three states that handed him the presidency — Michigan, Pennsylvania and Wisconsin — he won those voters by 21, 25 and 37 points, respectively. In each state, those voters made up about a fifth of the electorate.
Why? Why was Trump the preferred candidate among “I hate everyone” voters?
It’s hard to say. But if a substantial part of it was the fact that he was a relative newcomer to politics — the devil-you-don’t-know — that could be problematic in 2020.
It’s safe to assume that Trump and his allies will manage to make the eventual Democratic nominee fairly unpopular. If you think that’s not the case, consider the events of this week: Trump has managed to smear four Democratic representatives with a broad brush of anti-Americanism despite the obvious flimsiness of that charge.
If 2020 again comes down to Trump against an unpopular Democrat, though, will he still get the same benefit of the doubt? Or could the Democrat be the beneficiary of that motivation? It's impossible to say at this point.
But this group was essential to Trump’s win. Exit poll margins of error are wide, but minor shifts in the hate-everyone group — a few points in Michigan, six or seven in Pennsylvania — and Clinton would have won those states.
We come back to the original point, which was not “we know who will win next year.” It was, instead, that we don’t, that a lot of factors will come into play, including some we can’t predict. The lesson of 2016 is that weird things can and will happen, and that hand-wringing or confidence now can be perhaps only slightly better than wishing upon a shooting star.
Which, if you think of it, can’t hurt.