Doug Sosnik’s weekend op-ed provocatively titled “Trump is on track to win re-election” alarmed and annoyed many liberals of my acquaintance over the weekend. And in defense of annoyed people, large parts of the argument consist of sweeping assertions unbacked by evidence, like the idea that 2020 will see “other minor party candidates joining the presidential race.” One should also take any prognostications from the author of a June 2016 article declaring that “The 2016 election is already decided. History says Clinton wins” with a grain of salt.
That said, there is some real reason to believe that a betting person should regard Trump’s reelection as more likely than not. And Sosnik is correct to observe that Trump’s poor job approval numbers don’t tell the whole story.
Incumbents are usually reelected
There is going to be a presidential election on Tuesday November 3, 2020.
The previous evening, there will be a Monday Night Football matchup. I don’t know which teams will compete in that game or who will be playing for them. But I do know that the home team will probably win because what happens in pro football is that the home team usually wins.
By the same token, incumbent presidents usually win reelection. So do incumbent members of Congress. The extraordinarily high reelection rate of House members has a lot to do with gerrymandering, of course, but studies show that incumbency effects are very real. Bob Corker’s decision not to run for reelection in 2018 is good news for Democrats not because Tennessee is a particularly promising terrain for a Democratic Party pickup, but because an open seat is inherently easier to win than one held by an incumbent.
It would be a little hard to spin this basic point into a full-length op-ed, but it’s at the core of Sosnik’s argument and it’s correct. The incumbent usually wins.
There’s some reason to think Trump is unusually weak
If you want to believe that Trump will defy the historical pattern and lose, the best reason to do so is probably that he started out in an unusually weak situation for an incumbent. Sosnik writes that “Trump knows that gaining the support of a majority of voters in a presidential election is not a requirement; it’s simply an aspiration.”
This, however, has been true for a long time. Sosnik’s old boss, Bill Clinton, won in 1992 with just 43 percent of the vote — a significantly weaker showing than Trump’s 46 percent. That said, newly elected Clinton was broadly popular despite his weak popular vote total. FiveThirtyEight’s very useful comparative presidential approval tool shows that Clinton started out in January 1993 with a net approval rating (i.e., the number of people offering a positive assessment minus the number with a negative assessment) of 36 percent to Trump’s 3. George W. Bush lost the popular vote to Al Gore, but still started with a 32 percent net approval rating.
Other presidents came in even stronger. Ronald Reagan was at 38 percent net positive, Barack Obama was at 54 percent, and George H.W. Bush was at 49. What’s more, the fact that Obama came in with better numbers than the four presidents who preceded him shows that Trump’s rough start is not simply a consequence of long-term polarization trends.
Incumbents may normally win, in other words, but Trump is a lot less popular than an ordinary incumbent.
There’s good news for Trump
Sosnik’s best point is that looking at overall national job approval numbers likely underrates Trump’s electoral strength. As he writes, national polling does not “capture Trump’s strength in key battleground states.” Prognosticators who believe that one party or another has secured a long-term “lock” on the Electoral College are almost inevitably proven wrong. But we know that Trump’s Electoral College performance was far stronger in 2016 than his popular vote performance, and there’s little sign that the underlying demographic drivers of that advantage have vanished.
Approval polls also generally survey “all eligible voters, rather than registered or likely voters, which can underestimate Trump.”
In other words, just because most people are unhappy with Trump doesn’t mean that most people who will actually vote on Election Day in Pennsylvania will vote for his opponent. Most people don’t live in Pennsylvania and lots of people who do live in Pennsylvania don’t vote. All that said, to win reelection with a job approval rating that’s 18 points underwater sounds far-fetched. The genuine good news for Trump is that it’s totally possible these numbers will turn around.
Nothing matters this far out
Just 139 days into his presidency, Bill Clinton was 12 points underwater in job approval after a badly botched transition and a fumbling of the early legislative agenda that in some ways dwarfs Trump’s struggles. But he not only won reelection in 1996, he pretty much stomped Bob Dole, carrying states like Arizona and Kentucky on top of all the swing states he needed to win.
The secret to Clinton’s success was a lot more stuff happened over the next three and a half years.
Conversely, George H.W. Bush started out 49 points above water in job approval, was still about 47 points above water at this point in his presidency, peaked near the start of his second year at 77 points above water, and ended up unpopular and booted from office.
The problem, again, as Harold McMillan famously said, was “events, dear boy, events.”
Things happen and it matters.
The bottom line: unity matters a lot
Fundamentally, the odd thing about the 2016 general election is that both parties nominated someone who was a lot less popular than major party nominees normally are. Hillary Clinton’s arguments that Trump was unfit for office, as both a matter of temperament and character, were embraced by most voters.
Even in a critical state like Pennsylvania where Trump was supposedly more appealing locally than he is to a national audience, 56 percent said he was unqualified 60 percent said he was dishonest, and 62 percent said he had the wrong temperament for office. All of which helps explain why 52 percent of the state’s voters didn’t vote for him.
Clinton’s problem was that a chunk of that 52 percent didn’t vote for her either.
And that’s basically the problem Democrats face going forward. Trump unpopularity, even if it continues, only wins elections if people who don’t like Trump go vote for the Democratic nominee. If they fracture, then Trump could theoretically win even with a narrow base of support.
Sosnik sees this fracturing as part of a quasi-inevitable trend driven by the “lack of voters’ faith in both parties.” The actual history is considerably more complicated. Over 19 percent of the public voted third party (overwhelmingly for Ross Perot) in 1992. That number fell in 1996 even though Perot ran again and was joined in the field by Ralph Nader who drew about 0.7 percent.
Third-party voting overall was much less common in 2000 with Perot out of the field, but Nader’s support increased four-fold even though Al Gore was not less left-wing than Bill Clinton in any particularly obvious way. In 2004 and 2008, only about 1 percent voted third party, up to about 2 in 2012, and then surging to 5 or so in 2016 — far more than in other 21st-century elections but still far less than in 1992 or 1996.
I struggle to discern a clear pattern here that remotely corresponds with the (very real) long-term decline in the public’s confidence in the political parties as institutions.
The real bottom line seems to be that as with presidential approval itself, a lot hinges on events and the events themselves are highly uncertain. It’s certainly possible to imagine a world in which Democrats nominate a left-wing populist which inspires a centrist billionaire to run third party and a deeply unpopular Trump triumphs against a divided field. It’s also possible to imagine the steadily strengthening labor market evolving into a full employment boom over the next three years and Trump cruising to reelection on a wave of popularity. Or a recession could throw his approval ratings into the toilet and he gets crushed. Or there might be a nuclear war with North Korea.
In an uncertain world, the one thing we really know is the home team usually wins the game and the incumbent politician usually gets reelected.