Four more years? (ONE DAY to go edition!)

Well, here we are. The final update!

When we did the very first one of these, with eighteen months to go, I was arguing against the then fashionable Democratic peeing of the pants that Trump was inevitably going to win and we were toast. At one year out, I said there was still every chance we could do this, even though we didn’t know who “we” would be yet. At 6.5 months to go, I said that Biden’s chances were good, and Trump was in serious trouble. And last week, with one week to go, I concluded that Biden remained in an extremely strong position.

Well now, with one day to go, it’s put up or shut up time. Looking at the evidence, I think it’s overwhelmingly likely that Biden will win tomorrow (even if the result isn’t certain for a few days to a few weeks). Let’s take the evidence piece by piece.

Since last week’s update, the RCP national average has tightened somewhat, from +7.4% Biden to +6.5%:

If that’s all it’s moved in 6 days, it’s not likely to suddenly change in 24 hours, and if Biden is ahead by more than 6% nationally, he probably wins, no matter what chicanery the GOP pulls at the state level. Don’t just take RCP’s word for it, though. The 538 average is much more sophisticated, as it takes into account the historical accuracy of various pollsters, and their typical partisan lean. It has Biden up by 8.4%, down slightly from 9.2% six days ago:

It is of course state totals and the resulting electoral college total that is the final determinant. National averages are useful in setting a ballpark for state results, but to really get a sense, we need to look at the most likely swing states.

Last week, RCP had Biden ahead in 6 out of 6 swing states. It now shows him ahead in 5 out of 6, with some seeming movement towards Trump. However, RCP is also including some pollsters that have had a…questionable…habit of producing implausibly Republican-leaning numbers. 538, as mentioned, tries to adjust for things like this, and it shows Biden ahead in 6 of 6, and only Wisconsin having (pro-Biden) movement that’s outside of a tiny tick up or down that’s indistinguishable from statistical noise.

If one just takes the three states 538 has Biden ahead by more than a typical 3% margin of error and leaves everything else to Trump, this is the resulting map:

But, as mentioned, Biden also has leads in several of the other likeliest swing states, and in addition to those six, Georgia, Iowa, and Ohio are all widely thought to be in play. RCP’s “no toss-ups” map currently shows the following:

Taking one last look back at Biden’s position this year relative to Clinton’s in 2016 is also instructive:

To give a few highlights:

  • Biden’s lead is twice as large as Clinton’s at the same point.
  • On the eve of the 2016 election, 9.4% of voters were still undecided. Now, only 4.7% of voters are.
  • Biden is just shy of 51%. Clinton was below 47%.

In short, all the factors that made it more likely Trump could edge ahead of Clinton are not in play this year. And, indeed, 538’s forecast model now has Biden with a 90% chance of winning:

At this same point, Clinton was around 70%:

There’s a very important point to keep in mind about all the above polls as well. I mentioned this last week, but it bears repeating. These are not just leading indicators of how people may vote tomorrow. They’re also what people were saying as they were actually voting. As of this writing, early voting equals just shy of 70% of 2016 total votes:

The earliest votes started on 9/5, when North Carolina sent out its absentee ballots. Which means the above 70% of the 2016 vote all came in during a period in which Biden was never leading by less than 5.8%, and was frequently up by as much as 8% to 9%:

Finally, let’s check in one last time on a few other indicators. Trump’s popularity is more than 8 points underwater with one day to go until the election:

You probably don’t need anyone to tell you this, but that’s not a great place for an incumbent to be. And, indeed, his net popularity is well below any President who was going on to re-election in the past 50+ years. It’s more like Carter or Bush I when they were about to be defeated, or Johnson when he was so unpopular from Vietnam that he’d decided not to run again:

In terms of other elections, the RCP Senate no toss-ups map currently shows Democrats making a net gain of 3 seats, and tied in two others:

And 538 is projecting a 76% chance that Democrats re-capture the Senate:

The generic House preference poll shows Democrats with a more than 7% lead:

And 538’s model finds a 98% chance that Democrats keep the House, and even slightly increase their margin there:

We’re all understandably gun-shy after 2016. Could Trump win again, despite everything above? Yes, there are plausible scenarios. But every piece of evidence in view tells us that, with 24 hours to go, Biden remains in a very strong position, significantly stronger than Clinton in 2016. So take some deep breaths, remember to pace yourself tomorrow night and in the time to follow, but be encouraged.

And, tune in here in …a week? …two weeks? To see how the data measured up against the results!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s