I decided to go quiet on Twitter until after the election. But I had to document my thoughts in writing one last time before 11/3/2020.
Back in 2016 I was confident Hillary would win, but I did post two tweets running up to the election that belied my insecurities: here (11/4) and here (11/8). Now that we're a week from Election Day, I'd like to reflect for the record that I think this time around may be a blowout for Biden & what it means.
Two days after the 2016 election, I posted a bit of analysis on the numbers and takeaways. Mainly my point was that Trump's win was a razor-thin margin across MI + PA + WI & that demographics would make 2020 an even harder win ("hinge" election notwithstanding). I stand by my conclusions, although the certified numbers did shift somewhat (in Hillary's favor, actually). As many people have pointed out, third party voters & low turnout won't be the same factor this time around.
So what about now? TL;DR version: a big Biden win. And even better? I think we'll have a good idea on the night of the election because of Florida, Florida, Florida.
A few months back (this June) I posted this map as part of an argument to do away with the filibuster:
This map is simply Obama's 2008 win with three state changes: +GA, +AZ, and -IN. If you arrange the states by population, the largest 9 have a greater population than the top 41 (+DC): NC, GA, OH, IL, PA, NY, FL, TX, CA. Notice that six of these are very close this election! The number of people in those 9 states is 170.6M vs 161.9M in the other 41 (+DC).
If Biden is able to pull off such an amazing win, the 28 states he would be winning would represent 227.2M Americans from "blue states" while only 105.2M Americans would be living in "red states" ... less than half the population. If he miraculously adds TX, move 29.6M to the blue column (256.9M) & take it from the red column (75.7M). That would be an eye-popping 3.4x the population in blue vs red states. Talk about anti-majoritarian.
Check my numbers here in the first tab.
Why do I feel confident about such a big win?
538 puts Biden's polling average right now at 52.1% vs 43.0% Trump. On 11/1/2016, also a week before the election, 538 had it at 45.0% HRC vs 41.5% Trump. I know, I know, in the Electoral College system, national polls don't matter ... but they do. For the exact reason I laid out above: huge turnout = large number of people voting = reflect national polls closely & six battlegrounds are huge states.
On Election Day, Trump wound up at 46.4% vs HRC at 48.5% (~5.1% third party). So he gained about 3% and so did she. This extra six percent was easy to find because the polling average only contained 86.5% of voters. Not so this time around, when the polling average is at 95%. Really, Trump is still within the margin of error from where he was in 2016. But Biden is much higher than Clinton (7%!!!!).
Smarter people than myself have explained that the polls this time are not the same as in 2016. Although some of the margins for Biden in some of the polls are similar to Hillary's, one huge difference is that that while Hillary's polls often overlapped with Trump's & moved up & down, Biden's polling averages have been remarkably stable for months. Biden has also been at or above 50% for a long time, while Hillary's average never went above 45.8% (on Oct 25!). That matters.
Net favorability is another area (besides polling leads) where Biden has a clear and huge lead over Clinton. Whether fairly or not, the majority of people just didn't like her, while the majority of people do like Joe:
The graphic above is from The Economist, whose polling models give Trump only a 4% probability of success. The current 538 model gives him about 11%. So while this ain't zero -- as the probability of randomly guessing the outcome of a rolled six-sided die ~16.7% -- it's less than half of the probability that Silver gave Trump last time around.
In addition, many people in 2016 either voted for Trump -- or didn't vote for Clinton -- because they bought a hypothetical package he sold: that he was a successful businessman who would represent them in Washington. But 2016's populist promises & the reality of his record don't align well: his political agenda has been totally aligned with plutocratic Republican tax cuts for the wealthy & corporations. As a result, those who helped propel Trump to an unlikely win four years ago now have actions, and not just hot air, to judge him by.
Another bit of contrarianism: the outcome of the election may be clear sooner than you think.
Because of the stupidity of state-level laws that vary so much in a national election, some states are able to start counting mailed ballots now, while others (like PA) require waiting for them to come in on Election Day. Luckily a few states -- NC, AZ, and FL -- should have early returns, possibly even on election night itself, that may point towards a huge Biden win. Even if you can't call NC or FL within hours of polls closing, you can
start to see key precincts with huge turnout & figure out in a hurry
what the direction is going to be. If Biden is running up margins in NC, AZ, and FL, you can bet he's going to win PA, MI and WI. So my contrarian view is that NC will be called for Biden before we get ready for work on Wednesday morning, and this may be the writing on the wall for the end of Trumpism.
I feel encouraged in looking at the early voting returns in places like Florida. It was noted a few years back that a lot of Floridians were registering as "no party affiliation" (NPA). The map below shows the distribution of NPA voters in 2015 (latest I could find):
Now look at the early vote returns, courtesy the US Elections Project:
I am not sure how the NPA votes will break down, as a lot of "former Republicans" decided to go independent some years back (partly at the direction of Lou Dobbs). But a 6.4% edge in early voting is more than just alarming for Republicans in Florida, it's catastrophic. Of course they will spin it as, "Democrats are voting early, but won't show up on Election Day." It's quite possible that this is backwards, considering that in places like Florida Republican-leaning elderly residents have typically voted more by mail & in this pandemic you can bet even more will.
Of course I can't predict anything with certainty, but I've been obsessing over polls & returns for months & I think I can take a position that I'm willing to bet $ on at this point: Joe Biden is going to win, bigly. I'll stick with my map at the top of the screen, and I think TX will be very close.
The alternative -- four more years of sliding towards fascism -- is truly frightening.
