The FiveThirtyEight political analytics group has just unveiled its computer model for the 2020 presidential election. As you might expect from a casual glance at the polling, the model favors Democrat Joe Biden, but you might be surprised at how large a chance the model gives Trump of pulling off another upset victory.
The model, which is accessible publicly on the FiveThirtyEight website, differs from polls in that it does attempt to predict an outcome rather than simply recording a snapshot of public opinion at a given time. To do that, the creators of the model take polling as well as a number of other factors into account. For those interested in the nuts and bolts of the project, two of the model’s creators, Nate Silver and Galen Druke, discuss the assumptions of the model on a podcast that you can listen to here.
One important difference between the computer model and traditional polling is that the model takes the Electoral College into consideration. If you merely looked at national polling in 2016, you would have assumed that Hillary Clinton would emerge victorious… which she did in the popular vote. The rub was that Donald Trump eked out a slim victory in a handful of states that allowed him to sway the Electoral College result.
Once the model was built, FiveThirtyEight simulated the election 40,000 times and considered a variety of possible outcomes. The modeling shows that Joe Biden currently has a 72 percent chance of winning the Electoral College. While that reflects a substantial advantage, it does mean that Donald Trump can be expected to win in about one out of every four scenarios, which is a much greater chance than you might think from looking at raw polling data on sites like Real Clear Politics where Biden leads in the polling average by 7.5 points.
“It’s way too soon to count Trump out,” Nate Silver wrote in an accompanying article.
“While the polls have been stable so far this year, it’s still only August,” Silver said. “The debates and the conventions have yet to occur. Biden only named his running mate yesterday. And the campaign is being conducted amidst a pandemic the likes of which the United States has not seen in more than 100 years, which is also causing an unprecedented and volatile economy.”
“Nor has it been that uncommon, historically, for polls to shift fairly radically from mid-August until Election Day,” he added. “Furthermore, there are some reasons to think the election will tighten, and President Trump is likely to have an advantage in a close election because of the Electoral College.”
Currently, the model predicts that Biden will win 324 electoral votes to Donald Trump’s 214. The predicted popular vote outcome favors Biden by a 53-46 percent margin. Joe Biden is also estimated to have a 30 percent chance of a double-digit popular vote victory.
The model also identifies states to watch that are particularly close. These states will not surprise people who have been paying attention to polling thus far. FiveThirtyEight’s battleground states are include North Carolina, Ohio, Arizona, Georgia, Florida, Iowa, Texas, Pennsylvania, and Minnesota. Of those battlegrounds, Biden is forecast to win Minnesota, Pennsylvania, Florida, and Arizona, which would put him over the top in the Electoral College.
Inevitably, the question of 2016 comes up. People ask why polls and models should be trusted when they showed Hillary winning in 2016. The answer is that the failures of 2016 polling have been greatly exaggerated. National polling almost exactly matched the national popular vote. That did not decide the election, however.
State-level polling is always more suspect and variable, especially in close races. In 2016, 13 states were decided by less than five percentage points and state-level polling, in what the American Association for Public Opinion Research called a “historically bad year” in its post mortem of the election polling, had an average error of 5.1 points. This error was attributed to several factors including an actual movement of voters from Clinton to Trump in the last week before the election, an over-representation of college graduates, and the fact that late-deciders broke predominantly for Trump.
Despite the subsequent claims that “everyone” gave Hillary Clinton a 99 percent chance of winning the 2016 election, FiveThirtyEight’s model that year gave Clinton a 71 percent chance of winning, which left Trump with a 28 percent chance of an upset. Ironically, that is almost identical to current odds.
Trump supporters should not be too confident that the president will overcome the odds again, however. Silver points out that much Trump’s chance comes from the fact that the election is still a long way away.
“If we lie to our model and tell it that the election is going to be held today,” Silver says, “It spits out that Biden has a 93 percent chance of winning. In other words, a Trump victory would require a much bigger polling error than what we saw in 2016.”
“We should be clear: Trump’s current position in the polls is poor,” he adds, noting, “Biden is currently ahead in our polling averages in Florida, Wisconsin, Michigan, Pennsylvania, Arizona, Ohio and in the second congressional district in Nebraska — all places that Clinton lost in 2016.”
Trump’s hope is that between now and the election, something will happen to change public opinion and shift the polls in his favor. The pandemic and the associated economic downturn are the largest wild cards for the next few months.
One bit of good news for Trump supporters is that the Electoral College seems to favor Trump again. The model gives Biden a 71 percent chance of winning the Electoral College compared with his 81 percent chance of winning the popular vote.
While the outcome of the election looks dim for Donald Trump in mid-August, a lot can change in the next two-and-a-half months. That’s particularly true in 2020 where it seems that nothing is going normal. Watch the FiveThirtyEight model over the coming weeks to see which way the odds move.
Originally published on The Resurgent
