Recently I’ve seen a few pundits (mostly on the left) trying to prepare us all for a horrifying future: Trump will probably win in 2020, they say, because Presidents are usually re-elected. I’ve started to wonder: Is that true? Are Presidents usually re-elected?
There are two ways to look at this. First, we can put it in terms of the people who have been President. Here’s a one-paragraph summary of that version of American electoral history:
Washington was beloved, Adams was not, Jefferson was popular, Madison was popular enough, Monroe was very popular, JQA was good but lost, Jackson was bad but won, Van Buren was a schemer who lost, Harrison died, Tyler was an accident, Polk retired, Taylor died, Fillmore, Pierce, and Buchanan were technically President, Lincoln won and the racists murdered him, no one liked Johnson, Grant was beloved, Hayes sort of didn’t even win once, Garfield was murdered, Arthur was meh, Cleveland was chased out, Harrison was chased out, Cleveland got revenge, McKinley won and the anarchists murdered him, Teddy won, Teddy beat Taft without winning, Wilson won, Harding died, Coolidge was technically present, Hoover was loathed, FDR won won won, Truman defeated Dewey, Ike won, JFK was murdered, LBJ won and then gave up, Nixon won and destroyed himself, Ford never won anything, Carter got screwed, Reagan won, Bush barely happened, Clinton won, W won the second time, Obama won, and Trump got in by -3 million votes.
If you do the math, that amounts to 21 Presidents who have been re-elected, vs 22 who haven’t. (You might notice that this doesn’t add up to Trump’s favorite number, 45—that’s because Cleveland was only one actual person, and Trump hasn’t been eligible for re-election yet.) Of the people who weren’t re-elected, 9 lost in the election (like Carter), 5 were more or less primaried and didn’t even get to run (like John Tyler), 3 quit (like Polk), 3 died of natural causes (like FDR), and 2 were murdered (like JFK).
The most straightforward way to interpret this is that, by a small margin, Presidents usually don’t get re-elected, although the reasons are diverse and include a surprising amount of death. Sidenote: 8 out of 44 Presidents have died in office. That’s an 18% mortality rate! These days the most dangerous job in America is logging, and that fatality rate is about 0.1%—so historically, you’re way more likely to die if you become President. Still, healthcare is better than it was in 1850, and it seems like we’re out of that 100-year period where the President kept getting murdered. It might be fair to say that, barring death, re-election is somewhat likely.
At the same time, thinking of this in terms of people gives you some weird results. In some ways, what we really want to know is, sitting here in 2018, how likely is it that Trump gets re-elected? The data above isn’t great for that. Sitting in 1966, you had Lyndon Johnson, who counts as re-elected above, since he won in 1964, after serving one year—but he was eligible in 1968 and didn’t even run. And how do you think about Ford, who wasn’t technically eligible for “re-election”, never having been elected in the first place?
This leads us to the other way to think about this question, which is to put it in terms of elections. In other words, we can ask: For any given election year, was the guy in office chosen to serve again? There have been 58 Presidential elections in U.S. history, although we shouldn’t look at all of them. Obviously there was no one to re-elect in 1789, and in five cases the sitting President was legally barred from running again. In another eight cases, the President could have run again, but instead participated in the tradition of leaving after two terms.1 Though this wasn’t a hard and fast rule, it was a strong pressure against re-election, so let’s leave those elections out as well.
That leaves us with 44 elections in which the sitting President could realistically be re-elected. In those, re-election happened 22 times, exactly half—12 times the President didn’t run again, and in another 10 cases the President ran but lost. If we need a tiebreaker, we could count 1892 as a case when the President both was and wasn’t re-elected, since Ben Harrison lost to former President Grover Cleveland; that decides things for re-election by one (I was counting this as non-re-election, since the guy in office didn’t stay in office). But in any case, it’s tough to say the President is “usually” re-elected.
Of course, the people talking about the 2020 election probably don’t really care what happened with Cleveland and Harrison. I suspect they’re probably thinking, well, the last three guys got re-elected, and that gets you back about 25 years, so there you go. It’s fair to look at more recent elections, but this leads you to two problems. First, it’s tough to get enough data to draw any serious conclusions. The past 25 years only had three elections in which a President was eligible for re-election; not exactly enumerative induction.
Second, there’s no good way to know how far back to go. Starting at Clinton the data looks inarguable; go back one more guy and suddenly there’s a counterexample. Or say we start from the first year America had anything remotely resembling universal suffrage, with the passage of the 1965 Voting Rights Act. From there, five guys were re-elected and four weren’t (or, in the other terms, in nine eligible elections, there was a re-election five times). That’s hardly grounds for giving up hope in 2020. There’s no particular reason to pick any one cut-off date over another, so you can basically make whatever argument you want.
The long and short of it is that probably doesn’t make sense to talk about whether the President “usually” gets re-elected; it’s just sloppy thinking that gestures at history and empirical data without really engaging it. You’re better off looking at the factors that we know have a strong impact on election results—approval ratings, the state of the economy, gerrymandering effects, and so on.
Still, I think there are two pretty fair takeaways from all of this. First, if a President gets to the general election, he’s got a pretty solid chance of winning—that’s what happened in 22 out of 32 cases. But, second, it’s also pretty common for the President to fail, either through those 10/32 times the election goes against him, or because of the many factors that might keep him from running again. If he’s deeply unpopular, or he turns out to be in very poor health, or, just spitballing, the House discloses overwhelming evidence of Constitutional violations, criminal activity, and treason, it’s well within historical norms for the President to vacate the White House. The incumbency advantage is real, and we’ve learned that we should take the possibility of a Trump victory seriously. But we also shouldn’t resign ourselves to his winning. It wouldn’t be that unusual if he failed.
1. As FDR later showed, the traditional two-term limit was always a little more tenuous than it seemed; Washington obviously didn’t have to do it, Grant almost ran a third time, and Teddy Roosevelt eventually did run for a third term. I thought it was intuitive to count 1940 and 1944 as re-elections, even though I didn’t count, say, 1876 as a failed re-election for Grant (I just considered it an ineligible, traditional-two-term-limit year). I did count 1952 against re-election, since Truman chose not to run even though he was eligible, had FDR’s precedent immediately behind him, and hadn’t quite served two full terms. To me that seemed like Truman declining to run, although I think you could make a good case that this should also be considered an ineligible two-term-limit situation.↩