6/30/2020 Update: I have been bullish about Trump’s prospects from 2016, through the midterms, his impeachment, and even through the COVID-19 outbreak. In the early stages of the pandemic, Trump’s approval ticked up. And indeed, even today Trump continues to receive high marks for his handling of the economy, despite many Americans being out of work, facing foreclosure, etc.
However, many Trump voters, especially swing voters, seem to be turning on him — as a result of his grotesque response to the killing of George Floyd and the protests that followed. A trend that is most pronounced among white former Trump supporters. The president continues to hold his margins with minorities.
(8/6/2020 update: much more on this available here).
(11/2/20 update: still more on these dynamics available here.)
While there is still plenty of time go in this campaign — and the race will likely tighten depending on who Biden chooses as his VP (a decision far more consequential than usual given Biden’s age and health) — and as Biden is inevitably forced to emerge from his Coronavirus bunker rather than simply letting Trump destroy himself. However, as things stand, there seems to be a strong possibility that Trump could join Carter in being the only true cases since the Civil War where are party took power and didn’t manage to stay in power for at least 8 years.
However, what the investigation that follows definitely does suggest is that the landslide many seem to be expecting for Biden probably won’t come to pass. It’ll likely be much closer than most are expecting, and Trump could well pull off another upset.
“Everyone’s unhappy. Everyone’s ashamed. We all just got caught looking at somebody else’s page. Nothing every went quite exactly as we planned: Our ideas held no water, but we used them like a dam.”
Modest Mouse, “Missed the Boat,” We Were Dead Before The Ship Even Sank
In the lead-up to the 2016 election, Nate Silver offered Trump a 29% chance of taking the White House–and was widely disparaged for being far too generous to Trump!
Obviously, the Donald won. So what should we make of all the professional prognosticators who incessantly and near-unanimously assured us that it was impossible (or perhaps just ‘totally implausible’ if they were feeling charitable) that this outcome could come about?
Perhaps we should cut them some slack? After all, sometimes things happen that are legitimately improbable. Indeed, events such as these are often key drivers of history (which, arguably, makes them all the more important to “get right.” But that’s another matter). There’s a case to be made that Trump’s election was just this sort of phenomenon.
I don’t buy it.
Some Black Swans are events for which there was truly very little warning, or involve forces which are so complex and ill-understood that reliable prediction is impossible. Other Black Swans are merely events that are inconceivable given a certain set of assumptions, references or manners of thought. Trump’s election falls into this latter category.
Most analysts and pundits were absolutely convinced all the way until that last moment that Trump would not be able to win the Republican nomination. After he won the nomination, there was a lot of talk about lessons learned. However, almost immediately after Clinton secured her nomination it was depicted as virtually inevitable that she would win the general election—probably in a landslide (here, here, here for example).
Yet there was ample reason to believe things would go the other way. This is not an incident of hindsight bias: I realized early on that that Trump would likely not only win the Republican nomination, but the presidency as well. I predicted that he would beat Hillary Clinton explicitly, unequivocally and “on the record” for the first time in March 2016, and spent much of the remaining year urging Trump’s opposition to take his candidacy seriously and to better understand and respond to the factors driving his success. In vain.
After Trump won the general election, yet again there were avowals of lessons learned. But already narratives are emerging that 2018 and 2020 will be bloodbaths for the president and his party (if he even lasts that long in office). Rather than succumbing to this sort of wishful thinking, I decided to look for strong indicators as to how the race might go.
Polling this early in Trump’s administration seemed unlikely to be predictive—especially as the polls themselves proved somewhat unreliable in the 2016 cycle overall (particularly at the state level). Statistically speaking, it seemed the best way to start was to look at the base-rate for reelection in U.S. presidential races.
Establishing an Anchor
Franklin Roosevelt was elected to four consecutive terms as president of the United States, serving from the Great Depression to World War II. Out of concern some future leader might hold and consolidate power indefinitely, the 22nd Amendment was passed, limiting subsequent officeholders to a maximum of two terms. Eleven presidents have been elected since then: Truman, Eisenhower, JFK, LBJ, Nixon, Carter, Reagan, Bush I, Clinton, Bush II and Obama.
Eight administrations won a renewed mandate: Truman, Eisenhower, JFK/LBJ, Nixon, Reagan, Clinton, Bush II and Obama. Three presidents in this period stood for reelection and lost: Ford (who served out the rest of Nixon’s 2nd term), Carter, and Bush I. That is, the base-rate for presidential reelections is 8:3. Prima facie, one could assert that the current president would therefore have roughly a 72.7% chance of being reelected in 2020.
However, looking more closely at those who failed to win re-election, I realized that most were continuations of previous administrations: Ford followed after two terms of Nixon (and lost his bid to win a full term of his own). Bush I followed after two terms of Reagan (and was elected to one full term of his own). In the modern political era — in fact, going all the way back to the beginning of the 20th century — there was literally only one instance where there was a change of party in the White House that persisted for less than 8 years: the administration of Jimmy Carter from 1976-80.
| Party | Administration(s) | Tenure | Duration |
| Republican | McKinley/ Roosevelt I / Taft | 1897- 1913 | 16 years |
| Democrat | Wilson | 1913-21 | 8 years |
| Republican | Harding/ Coolidge/ Hoover | 1921-33 | 12 years |
| Democrat | FDR/ Truman | 1933-53 | 20 years |
| Republican | Eisenhower | 1953-61 | 8 years |
| Democrat | JFK/LBJ | 1961-69 | 8 years |
| Republican | Nixon/ Ford | 1969-77 | 8 years |
| Democrat | Carter | 1977-81 | 4 years |
| Republican | Reagan/ Bush I | 1981-93 | 12 years |
| Democrat | Clinton | 1993-2001 | 8 years |
| Republican | Bush II | 2001-09 | 8 years |
| Democrat | Obama | 2009-17 | 8 years |
| Republican | Trump | 2017-Present | TBD |
If we start the clock at FDR, the base-rate of reelection looks more like 8:1–meaning the ex ante likelihood of the Republicans winning reelection in 2020 would be about 88.8% (if we start with McKinley, at the turn of the 20th century, the base-rate goes up to 11:1 — implying a 92.7% likelihood for reelection in 2020).
But of course, it is not a simple task to infer the outcome of particular cases from raw-odds like these (i.e. comparing the reelection prospects of an abstract incumbent to those of Trump specifically). These odds are simply an anchor for analysis, to be revised down (or up) from there based on what we can discern about Trump’s likely situation in 2020.
But before we proceed to the next step, it is worth emphasizing how strong the historical trend is.
Let’s go go all the way back to the Civil War, and the formation of the modern political parties (Democrats, Republicans). Other than Jimmy Carter, there are only two exceptions to the norm of 8+ consecutive years of rule after a change of party… and both of those are only partial exceptions.
McKinley’s predecessor, Democrat Grover Cleveland, served 8 years as president — but his two terms were not consecutive (the only time in U.S. history where this happened). And he actually won the popular vote for all three elections between 1884 and 1896 (12 consecutive years).
The only other partial exception takes place in the immediate aftermath of the Civil War: Abraham Lincoln won two terms as President in the 1860 and 1864 elections. His fellow Republican Ulysses S. Grant won the presidency in the subsequent 1868 elections and served two terms. He was followed by fellow Republicans Rutherford Hayes, James Garfield, and Chester Arthur. However, Lincoln ran for reelection on a unity ticket with Democrat Andrew Johnson as his vice-president. And so, after Lincoln was assassinated, Johnson served out the rest of Lincoln’s term — until the slain president’s party regained power in the next election. Hence, it wasn’t 24 years of unbroken Republican rule, despite the fact that the GOP won all elections between 1860 and 1884.
And for that matter, although there was no Republican Party at the time (the closest approximation was the Whigs), and it was pre-Civil War, Lincoln was preceded by 8 consecutive years of Democratic reign: Franklin Pierce and James Buchanan (1853-61).
The takeaway should be clear: voters virtually always grant a party at least eight years in the White House. Jimmy Carter is a striking exception to the historical trend, and virtually the only exception to that trend, from the 1852 elections through the present. Indeed, Carter is literally the only exception to the rule, partial or otherwise, from 1896 through 2016 (i.e. the last 120 years). Let that sink in a moment, and then let’s proceed: