NewsWire: 11/3/2020
- With Joe Biden ahead in the polls, what does Donald Trump need to pull off an election victory? Nate Cohn of the NYT claims that the president needs "a polling error much bigger than 2016" if he is to hold on to the White House. (The New York Times)
- NH: If you look at the polls, Biden has a significant lead over Trump. RCP has Biden up by 6.7% nationally. PredictIt, a political futures market, has Biden winning 64¢ to Trump's 42¢. 538's election model gives Biden an 89% chance of winning. And The Economist model gives Biden a 96% chance of winning.
- But all these polls could be wrong. Or, more to the point, they could be right and Trump could still win. After all, Nate Silver (of 538) is still giving Trump 10%. That would be a surprise--but not inconceivable. Silver, btw, also says there's another 30% chance that Biden will win, but it will take longer than election night to call, and another 60% chance he will be declared winner by 3 am tomorrow.
- To say a Trump victory is conceivable, on the other hand, is not to say that it's likely.
- To begin with, contrary to urban legend, the polls in 2016 weren't as off base as many suppose. They correctly showed Clinton winning the popular vote, and many stayed within their margin of error.
- A Pew Research analysis found that national polls in 2016 were only off by an average of 1 to 2 percentage points. (Bottom line: Most of them had Clinton up by 3 or 4 points, and she actually won by 2.) They were far better than polls in most presidential races on record. The problems mainly stemmed from a lack of quality polls on the state level and an underrepresentation of whites without a college degree. It was also fortunate for Trump that the biggest polling misses by state just happened to be in the four or five states where he won by very slim margins.
- We also hear about "shy Trump voters" and how they could be generating an underestimate of Trump's support. But most tests of such shyness has failed to detect much evidence of it. The theory isn't helped by a recent finding that the single biggest source of new support for Biden in midwestern battleground states are voters who tell surveyors that they are "switching" from Trump in 2016 to Biden in 2020. It's hard to argue that there's anything shy about someone who admits to switching.
- With all this in mind, let's look at the current RCP 7-day average going into election day. It shows Biden ahead of Trump by 6.8 points. That's 3.5 points more than Clinton was leading Trump on election eve. What's more, Biden's lead--while narrowing a bit over the past two weeks--has been quite stable compared to Clinton's lead in 2016. That stability comes from a paucity of undecided voters and very little voter appetite this time around for any third-party candidate.
- So let's move on to Nate Cohn, who asks in this piece: If the current polls are as wrong as they were in 2016, would Trump win? His answer is a flat "no." Assuming the same level of error as 2016, Biden would still win with 335 electoral votes. If it goes the other way, as in 2012, and the polls underestimate Democratic candidate's margin, then Biden wins in a landslide with 412 electoral votes and he even beats Trump in Iowa, Ohio, and Texas.
- So what is Trump's path to victory? Well, IMO, there's only a narrow path. Three things would all have to happen.
- First, the national polls would have to be way off, enabling Trump to beat them by a larger margin than in 2016.
- And second, the polls in the battleground states once again would have to be even more tilted against Trump relative to the national average than in 2016. This will harder this time around because there are a lot more battleground states. If, as seems very likely, both Michigan and Wisconsin go to Biden, Trump needs to draw even with Biden or better in almost every other winnable state.
- And third, assuming this pushes just enough states into the zone of contestability, Trump needs to win all his legal battles. Maybe he wins them outright. Or maybe he just succeeds in preventing one or more state governors from clearly certifying only one panel of Electors before the safe harbor day (December 8). Trump could thereby get those Electoral votes thrown out in a (likely highly contentious) joint session of Congress on January 6, 2021.
- If you want a highly in-depth look at the election, watch my latest webcast with JT Taylor, Hedgeye's Chief Political Strategist. See "Election 2020: States of Play." We go through all the possible scenarios and what you need to know for election day.
