We won’t really know who is going to win the presidential election until November 4, and possibly not until the Electors cast their ballots on December 14. But if we think about the structural factors in conjunction with poll results we can gain a much clearer idea of the possibilities.
I start with the principle that however hard or easy it may be to chose one candidate or another in the first place, once we have decided and declared our decision — even if only to ourselves — we grow very unlikely to switch to another candidate. And the starker the contrast between the candidates appears to be the less likely we are to switch. In today’s highly polarized environment, switching from Trump to Biden, or the other way, is all but unthinkable.
The next thing to think of is the polls. A lot of people feel badly burnt by the 2016 polls and are inclined to ignore them in 2020. That’s pretty unrealistic. The polls in 2016 were as good or bad as they have ever been. The problem was the unthoughtful meanings people tried to impose on the numbers. The polls as a whole said that Clinton was ahead in the popular vote, and indeed she was. If you analyzed the polls well, you came up with an estimate of her popular-vote lead that was very close to the actual results. But of course U.S. presidential elections are not decided by national-level popular vote. For reasons that made sense in the 1780s, “electoral votes” were apportioned to the States by a formula that made voters in, say, Wyoming a great deal more equal than those in California. Since it is written into the Constitution it is far from easy to change, much as many people would like to. Until then, it is necessary to look at the polls State-by-State to translate human votes into electoral votes.
One of the things that led to some of the surprise in 2016 was that a lot of voters remained undecided until very late in the process. Both of the major candidates were widely disliked and distrusted, making the choice unusually difficult for many people. At the very end, polls were showing 13% or so undecided, or planning to vote for third-party candidates. (People who say they are going to vote third-party are often not deeply reluctant to switch to main-party candidates, experience shows.) At the last minute, even on Election Day itself, quite a lot of these people decided to vote for Trump — well over half of those who voted at all chose Trump, which was enough to flip several critical States into the Trump column. The lesson is, watch that undecided percentage.
As I write, the election is just over six weeks away, and things are not looking terribly bright for Trump. It’s not that he’s really all that far down in the polls, but the patterns present problems. He remains popular in States which are relatively overrepresented in electoral votes, as he was in 2016, and might win the electoral vote even if he lost the popular vote by 3% or even 4%. (He lost it by 2.1% in 2016.) But Nate Silver’s FiveThirtyEight.com presently projects that Biden/Harris will win 331 electoral votes, meaning that Trump/Pence would need to take away 62 of these in order to turn the tables. Most other serious analysts are pointing in the same direction — see the Cook Report Electoral College forecast, for example. If Trump & Pence held on in all the States where they currently lead (by Silver’s count) and also win in NC, FL, and PA, they they could get just enough electoral votes to win. So let’s look at these States.
North Carolina (15 electoral votes). Biden’s lead over Trump in NC is quite narrow, little more than 1%, although it has been remarkably consistent throughout the whole campaign. With 6% undecided, gaining that extra 1.5% does not seem totally out of Trump’s reach, but something is going to have to change for it to happen. Of course North Carolina has an especially rich history of election chicanery. One way or another, it’s really pretty early to count on NC going for Trump or for Biden.
Florida (29 electoral votes). Biden has led since early April and currently has a margin of 2.0%, with 5.4% undecided. In order to win, Trump has to pick up a very large fraction of the undecideds — two out of five at very minimum (with none voting for Biden), and probably a lot more. Election chicanery is far from unknown in the Sunshine State, and Gov. Ron DeSantis is very beholden to Trump, but this still looks like a pretty big lift.
Pennsylvania (20 electoral votes). The Keystone State could very, very easily be the one that decides the election. Biden has led significantly throughout the entire campaign in PA and right now polls at 49.7%. Trump is at 45.1%, so undecideds amount to 5.3%. That is to say that virtually all of the undecideds have to go for Trump if he is to win. Watch for lots of action in Pennsylvania, as well as North Carolina and Florida.
I want to draw attention to the consistency of the Biden lead in these States. That’s historically unusual, and quite different from 2016, and it suggests that a great many people made up their minds — generally as to whether to vote for Trump or against him — a very long time ago, and are not very likely to switch votes. It really is likely to be up to the undecideds, and Trump has to convince an awful lot of them if he is to win.
The first presidential debate comes up on September 29. The first debates have generally drawn big audiences and generate a lot of comment. But whether they or the subsequent debates exercise much real influence is far from clear.
That gets us to the October Surprise. It’s not even October yet and already we’ve had a major surprise in the death of Justice Ginsberg. As we see, Trump and Mitch McConnell at once decided to attempt to exploit this for advantage. It’s not clear what the outcome will be in terms of confirming or not confirming a new justice, but it is hard for me to see the Republicans extracting much electoral advantage from the spectacle in either case.
We can scarcely doubt that Trump has Roger Stone and other seasoned dirty-tricksters, not to mention executive branch loyalists very ready to bend their offices to serve his political ends, working feverishly to come up with something. In 2016, after all, James Comey, violating FBI and DoJ policy, announced what turned out to be a baseless innuendo against Hillary Clinton on October 28 which very probably won the election for Trump. Surely Trump’s men will come up with more dramatic (and probably even more meretricious) accusations to hurl against Biden.
But the unstable conditions that gave Comey’s letter such effect in 2016 are very far from what we see today. Public confidence not only in Trump but in anyone even remotely associated with him is very low. Having been badly used in 2016, the major news media are more cautious. Anything the Trumpists come up with will be seized upon by Fox News and other elements of the far-right media, of course, but it is questionable how much they influence the views of the undecided voters. We cannot entirely dismiss the impact of any such surprises, but it is far from clear that they can be decisive.
Moreover, as Bob Woodward has recently demonstrated, surprises need not be favorable to Trump. And more books of a potentially explosive nature are due out in October. Yet although Woodward has Trump on tape acknowledging that he cynically misled the public on a life-and-death issue, the response in the polls has been scarcely visible. This too speaks to the difficulties of changing the balance in this election, even with dramatic and true revelations.
Finally, there is the question of the influence of Vladimir Putin. There is little doubt has set his troll troops to spreading disinformation and dismay in an effort to aid Trump to stay in the Oval Office. And indications seem to be that Mark Zuckerberg and Facebook are very open to their efforts. But even crediting Messrs. Putin and Zuckerberg with the very worst possible intentions, I don’t see the Web campaign as at all decisive. It did not seem to have much visible effect in 2016 and people have grown more skeptical in the meantime. How many of the undecideds get a lot of their news and views from the social media, after all?
But what about Internet sabotage of the voting process? There are a lot of structural reasons why that is quite difficult in the United States, and the State and local election authorities all seem to be on the alert. Moreover, U.S. Cyber Command is now very actively involved, including preemptively striking at hostile Internet sites.
Finally, voting is underway already. It seems likely that a very substantial portion of ballots will be cast well before Election Day, particularly those for Biden & Harris. Thus what we see in the polls today is in substantial part a measure of voting reality. Undecided voters are not going to vote unless and until they decide, of course, but whatever window there may have been for flipping Biden voters is closing fast.
My bottom line is that there is still room for Trump to win, but only narrow room. Absent something quite unexpected and dramatic, Biden will win the election.
Great analysis!