The hard data compels me to predict that Joe Biden will win the presidency. The polling averages, both nationally and in the key battleground states, show a lead that is both more durable and perhaps more comfortable than Hillary Clinton’s four years ago. The down-ballot polling looks better for Democrats and seems to undercut the shy Donald Trump voter theory. The early voting trends appear favorable to Biden.
Then there is just the gut-level feel of the election. Trump won in 2016 by receiving less than 80,000 more votes than Clinton in the Rust Belt states that swung the Electoral College with significant third-party candidates on the ballot. Remove those candidates and add in the pandemic, the resultant economic downturn, and the various racial justice protests, and you erase more of his support than the movement of some Evan McMullin (and maybe Gary Johnson) voters back into the Trump column can possibly compensate for.
But unlike most believers in this conventional wisdom, I don’t dismiss the case for a Trump win. I find many of the arguments plausible. I just don’t see enough evidence to make them myself. Nevertheless, just as Biden is within striking distance of the biggest Electoral College landslide since at least 1988, all it takes is a few points for Trump to hold the Sunbelt, Pennsylvania — and the presidency.
Then there is the small matter of polling in a pandemic, especially when one candidate’s supporters are substantially more risk-averse than the other’s. Could that swing Election Day turnout as much as it does campaign event turnout, even to a degree early voting can’t protect against? I haven’t seen polling misfires in any of the post-COVID-19 primaries or special elections serious enough to make me predict it. But it is not a crazy theory.
It is probably too much for Trump to hope the conditions of 2016 can repeat themselves. But because 2016 did happen, Biden can’t dismiss the possibility either. So I bet Biden, but cautiously.