Donald Trump could be the worst loser in recent election history
The magic number is 37.4%, and some polls place it below that bar.
Public humiliation is probably Donald Trump's greatest fear, and chances are the one that awaits it is historic. However, these polls, conducted before the reopening of the Hillary Clinton email investigation, are to be taken in hindsight, as they could be upset.
The most unpopular US presidential candidate in modern history is George McGovern, whose 1972 campaign against Richard Nixon resulted in a small score of 37.4% of the vote. There is at least one commonality between the political landscape then and today: while it was not as straightforward as the Republican Party's shame for Trump in 2016, the Democratic establishment of the he era had abandoned McGovern, then darling of the most leftist activists.
The political consequences have been immense. McGovern continues to haunt Democrats today. Hillary Clinton worked for him in Texas at the time and often uses her shattering defeat as a warning to progressives who want to go too far at the risk of ending up with a Nixon in the White House. The Democratic Party transformed in the years that followed, culminating in the centrist presidency of Bill Clinton.
Just as the left was buried in 1972, the United States now has the opportunity to bury not only Donald Trump, but also the racist and misogynistic politics he represents. Complete rejection, with the eyes of the whole world on the country, could send a clear message: Donald Trump is not America.
23% chance of making the worst score in presidential history
The Huffington Post did the math, and today Trump has a 23 percent chance of scoring McGovern below on election day.
Trump is lagging behind Clinton in the presidential race. According to the Huff Post poll aggregator, he is 7.7 points behind Clinton in an opposition two, and 6.5 points behind if you include the smaller candidates.
In most polls, Trump has between 38 and 42 percent of voting intentions. In those that include the small candidates, he's 23% more likely to do worse than McGovern. His scores are higher in duels against Clinton. If you consider these to be closer to reality, then he has a 4% chance of doing worse than McGovern.
People who don't want to vote for Trump or for Clinton play an important role in this Biggest Loser Crusade. Any ballot in the ballot box that is not in Trump's name helps decrease his percentage of votes, as it increases the total number of votes cast. Voters of the Constitutionalist Party (a nationalist party that won 0.09% of the vote in 2012), or those who slip in a ballot with the name of a bogus candidate, do have a role to play this year.
To arrive at this result, we took into account all the polls conducted entirely during the 50 days before the November 8 election. In other words, the polls that ended on September 19. According to an analysis by political scientists Robert Erikson and Christopher Wlezien covering the elections from 1952 to 2008, polls that are taken 50 days before the election predict approximately 75% of the variance of the real ballot. The predictive power of polls increases by up to 95% when they take place on election day.
The deadline of September 19 gives us, this Thursday, October 27, 87 polls that include small candidates, and 65 polls that concern a Trump / Clinton duel. Trump has between 37 and 49% of those polled in his favor in the duels. When we include the small candidates, the billionaire collects between 36 and 44% of the voting intentions.
Polls less and less vague as D-Day approaches
Of course, the closer the election gets, the more the polls are a good indication of the end result, so we weighted Trump's scores in the polls based on their proximity to the poll. These coefficients were measured using the approximate variance explained by the equivalent surveys in the analysis of Erikson and Wlezien. Thus, the polls carried out at 50 days (September 19) have a coefficient of 0.75, then this coefficient increases by 0.004 for each day that approaches the election.
This weighting accounts for the greater or lesser uncertainty depending on the day on which the survey is carried out. This factor is important. We know there is an element of uncertainty in polling, and the further away from election day, the harder it is to predict the end result. The time-weighted average of the percentage of the electorate supporting Trump is used as a benchmark for where the billionaire stands.
The weighted average is 39.2% for Trump in polls that include small candidates. The weighted standard deviation, which shows the variation between different surveys, is 2.4 points. To calculate the probability of Trump falling below 37.4%, all you need to do is do a simple statistical operation.
It is assumed that the margin of error goes both ways, above or below current estimates: Trump could just as easily gain support among the undecided as he lose votes among those who say they support him (if for example a new scandal breaks out). We thus distribute the scenarios fairly, estimating that Trump has a 50% chance of scoring above the poll average, and a 50% chance of scoring below.
The result we are interested in, 37.4%, is 1.8 points lower than the 39.2% average, which means that Trump has a 23% chance of getting 37.4% of the vote or less.
If we do the same calculation with the polls that only take into account Trump and Clinton, then Trump has only a 4% chance of getting 37.4% of the vote or less. This is explained by the fact that the weighted average of the Republican candidate in these polls is higher (42.4% with a weighted standard deviation of 2.7 points).
Studies show that the end result of the election mostly lies between the estimates of polls that include small candidates, and those of polls with only the top two contenders. If we apply this general rule, then Trump has a 10-15% chance of getting 37.4% of the vote or less.
This is of course an approximation. But it shows that Trump has a non-negligible chance of becoming the worst loser in recent American presidential history.
