Now that Trump has essentially won the GOP nomination, those of us who live in the U.S. have to confront a deeply troubling reality: At some point in the near future, tens of millions of our fellow Americans will have voted in favor of Donald Trump becoming President. This seems destined to go down in history as a fascinating combination of shame and idiocy.1 But just how bad will it be? How many of us are going to officially register our opinion that Donald Trump is not only acceptable but also preferable to another human being as a candidate for the world’s most powerful job?
In this post I attempt to find the Trump Minimum, the absolute best case scenario for this nation as it attempts to live with itself from 2017 on.2 The question: What is the lowest possible number of people who will wind up voting for Trump?
I’m ignoring two possibilities here. First, that Trump won’t run—either he’ll suddenly realize that he might actually have to do the job, and panic and quit; or he’ll run out of money because he has been a secret poor person this whole time; or Paul Ryan will decide that he’d like “history” (one year from now and after) to remember him kindly and will orchestrate a convention coup. In that case Trump will only have his primary votes; these promise to be pretty substantial, already topping 10 million as of this Washington Post article from April. I’m not sure primary votes matter in quite the same way, though. The stakes are just lower in those contests; some people might be voting tactically without hoping for the candidate to win, and others might just be gambling.
The second possibility is that Trump wins. Recent polls do show the field narrowing. Personally, I don’t think this is very likely.3 But if he does win, we’re going to have way more important things to worry about. (Also, I suspect that if he does win we might be more angry at the people who failed to vote against him—who knew he was bad and let it happen anyway. But that’s just a guess.)
This is all going to be pretty straightforward. I take a margin of victory for the Democrat—I’m just going to say Clinton from here on out, since she’s more likely to win the nomination—along with the third party vote and use that to figure out what percentage of the popular vote Trump receives.4 I’m going to ignore the electoral college as well as demographic breakdowns; both are important and interesting, and the latter affects the overall popular vote, but I really just want big-picture numbers here. I’ll leave the story behind the numbers to the reader’s imagination.
I also estimate overall turnout, which tells me how many people are voting in the first place. I found it pretty tough to find a good source for how many eligible voters there are in the USA, so I used this reliable-seeming Wikipedia page on Voter Turnout in U.S. Presidential Elections, which shows this trend:
That’s not the world’s worst trend line, so I just used it to predict that there will be 240 million eligible voters this year. That feels like kind of a strange method, and definitely not like very good social science, but there were 235 million eligible voters last time, so it seems like as good a guess as any.
We’ll start with a pretty plausible, and therefore slightly depressing scenario. The Real Clear Politics average of polls has the race essentially tied as of this writing, so we’ll use a previous close election as our Sad Baseline: In 1960 JFK defeated Nixon by just .17% in the popular vote. That’s perfect for a sad winning margin.
Since 1932, which is when my data starts for this, turnout has ranged from 49% (in 1996) to 62.8% (in 1960—so not only was the result a dead heat, but tons of people weighed in on it). In the past few elections turnout has hovered in the low to mid 50’s. Since this is the Sad Baseline, we’ll assume decent turnout of 55% (thus boosting the raw numbers of Trump voters). And finally, we’ll assume that third party performance is just under the historical median—call it 2%. All together, this gives us:
That’s over 64 million votes for Donald Trump. Thanks to historic population growth, only Barack Obama has ever received more votes in a real Presidential election (though he did do it twice).
Those assumptions were about as pessimistic as possible—good for figuring out the top of the range, but otherwise counter to the spirit of the Trump Minimum. So let’s consider a Plausible Good Baseline. In 2008 Obama beat McCain by 7.27%. That might be about as strong a margin as you can hope for in an era this partisan, so let’s steal that. Let’s also gamble that the historically bad favorability ratings for both candidates will depress turnout to a tie with its worst level in the last 80 years—49%. Those adjustments to the previous scenario lead you to just over 53 million Trump voters.
We’ve still got one powerful lever, though: Third Parties. These have had a pretty substantial spoiler effect in U.S. History; it was only 24 years ago that Ross Perot pulled 19% of the vote. So far there’s no indication that anyone will come close to matching that, but there’s some evidence that Libertarian candidate Gary Johnson could reach as much as 10% of the vote. If we bump our third party cut up to that, and keep everything else the same, we arrive at a Plausible Gary Scenario wherein Trump receives just 49 million votes.
All of this is well and good, but what if we ratchet the numbers up to a historic disgrace? Is it really so implausible that Trump will do or say something that loses him the votes of 9 out of 10 women? Or that he will quit the race and return to it the next day, on multiple occasions? Or insist that Donald Sterling be his Vice President? In hypothetical times like these, we really ought to turn to the biggest margin of victory in modern U.S. Presidential election history: Warren G. Harding’s 26.17% blowout of James Cox.5
Let’s also start thinking outside the box about voter turnout. Sure, 49% is bad, but what about the examples of other nations? Surely some of them care even less than us. I looked into it and, no, they really don’t, at least not in countries that are doing pretty well.6 Among OECD nations, only Japan, Chile, and Switzerland are more apathetic about voting. Only 40% of Switzerland’s voting age population votes, so let’s just take that as a worst-case-scenario number for U.S. turnout, too.
Finally, we’ll ratchet Gary up to 15%. Why not? After all, in this scenario Trump keeps quitting and promoting a man who was banned from the NBA.
With these parameters, we have reached the Trump Minimum, the absolute lowest number of votes that he could reach if we hit historically plausible extremes. The final count: 28 million votes.
So there you have it. Donald Trump is virtually guaranteed to get at least 28 million votes in November. A minimum of ~30 million people living in this country will choose to place Donald Trump in charge of maintaining the nuclear arsenal, repairing our broken justice system, and engaging in diplomacy with leaders who often are not old white men.
I’ll admit I wanted that number to be lower. I even played around with one last scheme: The Return of Teddy Scenario. In this one Teddy Roosevelt returns from the dead along with his former opponent Eugene Debs, and together they replicate their record-setting third-party performances in the 1912 election to steal 33.4% of the vote. Gary Johnson still runs (nothing has stopped him so far in real life), though he only adds 10% to the third-party haul, since some of his coalition prefers the Bull Moose charisma. Most voters are turned off by the prospect of voting for a reanimated dead person, so turnout dips to 30%. And as a hyper-masculine old-money white man with century-old values, Teddy pulls disproportionately from the Republican vote; Debs does pull some Bernie supporters out of Clinton’s coalition, but her margin still improves to an even 30%. In the Return of Teddy Scenario, which is not likely, Trump receives just 13.3% of the vote, losing to both Clinton and a long dead park enthusiast. But he still pulls 11 million votes.
There’s just no way around it: The man is going to get a lot of votes. Realistically closer to 60 million, but at the very least about 30 million. We’re all going to know someone who voted for him; we’re all going to go into future elections with hard evidence that people who like him are out there voting again. But there’s still the possibility that, like Barry Goldwater or Walter Mondale, he’ll lose badly enough that we can think of ourselves as the country that overwhelmingly rejected him. We just have to hope he’s less like Richard Nixon and more like James Cox—his electoral performance remembered mainly for how thoroughly he was defeated.
1. I used to say this about re-electing George W. Bush, too, so it’s possible that I’m either overreacting or underestimating what’s next. Also, I was originally calling it one of our most shameful and idiotic moments, but then I thought about things that happened before 1950. ↩
2. To be fair, we often don’t seem to have a problem hanging around with evidence of our worst moral failures. Strom Thurmond, who ran for President in 1948 on a pro-Jim Crow platform, only left office in 2003—and he wasn’t even voted out; he just died. ↩
3. Anything can happen, and major party nominees always have a decent enough shot at the White House that it makes sense to worry about the dangerous ones. But speaking purely subjectively, I just don’t see how he overcomes demographic reality. Women and people of color hate him even more than they hated Romney and McCain, who both lost. The theory behind his winning rests on activating enough “missing white voters” (read: racists) to overcome this, but I don’t see how enough of them are A) still alive, B) eager to vote now, despite having sat out two elections featuring a black candidate. ↩
4. I don’t allocate the third party vote to one party or another, but it shouldn’t really matter for these purposes. If you thought a Libertarian, for instance, would take many more voters from Trump than from Clinton, you could just change the Clinton margin of victory; but I’m just making up that margin in the first place, and why complicate a made up number? We don’t care why people are or aren’t voting for Trump—we just care how many of them there are. ↩
5. I don’t know why the 1920 election in particular. It’s interesting that it’s the first one in which women could vote, and the second biggest margin was in the very next election. Maybe only one party appealed to women voters? Three other things I want to put in this footnote: 1. I’m saying “modern” elections because my data only goes back to the 1820’s. Before that Monroe and Washington essentially ran unopposed, which screws up projects like this one. 2. Kind of funny that the VP on Cox’s ticket was FDR—guess he made up for it later. 3. I thought of Donald Sterling as a joke and was immediately convinced it was plausible. What better choice for Donald than, essentially, himself: a racist real estate billionaire named Donald. ↩
6. I submit this as a replacement phrase for “developed countries”. ↩