2020 election predictions

November 1, 2020 2020 election predictions

PRESIDENCY

NATIONAL POPULAR VOTE: Biden has a 7.2 pt lead according to Real Clear Politics (RCP) (more according to some more Dem-leaning poll aggregators).  Trump did about 1 pt better than the national polls in 2016, and I don’t see any reason to believe that this will be different.    So I am going to guess that Biden will win the national popular vote by around 6 – say 53-47 minus third party votes.  This means that the electoral map will look fairly similar to 2016 but that Biden might pick up the swing states he needs to win. Trump won half a dozen states by less than 2 pts (NC, Fla, Pa, Mi, Wi, Az) and Biden needs about half of them to win- fewer if he wins another state where he is polling decently.

NORTHEAST:  The only state that seems to at all be close is Pennsylvania; Biden seems to have a significant lead everywhere else, even in New Hampshire which was close in 2016.

So what happens in Pennsylvania?  The case for Biden: 1) the Democrats only lost by 1 pt or so in 2016, and so if you assume a 2 pt nationwide shift he wins; 2) he seems to be leading in most polls by 4 pts so if the polls are as off as in 2016 (2.6 pts) he still wins (but by only a point or so so that’s not exactly a comfortable margin); 3) polls are trying to weight for education to avoid underrepresenting the white working class, so they may be less inaccurate than in 2016.

On the other hand, 1) voter registration has been trending Republican for the past several years; 2) Biden leads by less than in Michigan and Wisconsin (the two other Rust Belt states that unexpectedly voted for Trump in ’16); 3) poll data so far doesn’t seem to reflect vote-by-mail totals. 

It looks like among mail voters, Democrats lead Republicans around 3-1 (66 percent D, 23 percent R, 10 percent independent, so overall about 71-73 percent Dem assuming a roughly equal number of people cross party lines and that Inds are slightly Dem).  But the polls I have seen divide voters into people who have already voted and those who have not voted yet, and some of them they show Rs getting 20 pct or less percent in the “already voted” category.  For example, this poll says Biden leads by 57 percent (78-21) among already-voted voers. Unless registered Republicans are deserting Trump in large numbers, that suggests polls are not picking up on a significant number of R mail voters.

On balance I am calling Pa for Biden but very, very reluctantly. If Biden was leading by less in the popular vote I’d probably say Trump carries it.

THE SOUTH:  Most of the South is pretty safe Republican, and Virginia is safe Democrat.  A few states are worth discussing:

North Carolina- polls show a close race.  Vote by mail totals show that Democrats lead Republicans in early voting but their edge is comparable to their overall voter registration lead.  Given that Repubs are more likely to vote on election day, this strikes me as some evidence that Trump will prevail.

On the other hand, every poll in RCP but Rasmussen (which tends to favor Rs) shows Biden with a slight lead. And every Senate poll shows the Democrat with a slight lead. Why does this matter? If there were a significant number of Trump voters lying to pollsters they would feel more comfortable admitting they were supporting the low-profile Republican Sen. Thom Tillis than they would be to admit they supported Trump. So Tillis’s failure to run ahead of Trump suggests people are telling the truth. (Does this argument reflect people simply refusing to respond to polls? No. But polls are often weighed by age, education etc to eliminate these biases). I am reluctantly calling NC for Biden but not with a lot of confidence.

Georgia- Another too close to call state.  Because polls are mixed and Ga. has voted Republican in the last several elections (even in 2018 with a terrible gubernatorial candidate), my instinct is that Trump carries it.  But if Biden does better than I expect Ga. will be one of the dominoes that are likely to fall.

Florida- Fla is so Republican that it threw out a Democratic Senator in 2018 (although not by a large margin). Also, the Rs ran a few pts ahead of polls that year. Given that fact, I think Fla is pretty likely to go for Trump.

Texas-  Texas has been a solidly Republican state for the last 20 years.  Also, the RCP poll average still favors Trump slightly.  So I think Trump holds it.  But if Biden wins by, say, 9 points instead of 7, Texas is another domino that could fall. I have to admit, huge voter turnout makes Texas a wild card though.

MIDWEST- Illinois is safe D, and Indiana, Missouri, ND, SD, Ks, Nebraska and Oklahoma are safe R.

Ohio- RCP average shows Ohio a tie.  Given its history of supporting Trump I think Trump probably wins- but as with Texas, if Biden does just a little better than I expect he can take it.

Michigan- Trump carried this state in 2016 but Biden seems to have a pretty consistent lead here.  (One recent poll, Trafalgar, shows Trump winning but that seems to be a pretty consistently pro-R poll- also, the poll says about a quarter of voters on each side are switching parties which seems highly unlikely, so I question its usefulness) I say Biden flips it.

Wisconsin- The polls pretty consistently show Biden leading in Wisconsin.  On the other hand, the polls were more wrong there in 2016 than anywhere else; Trump ran 7 pts ahead of the RCP poll average.  (Today’s RCP poll average shows about a 5 pt lead for Biden.)  But 2016 polls really didn’t weigh for education and thus underestimated white working class turnout.  It seems to me that this year’s polls are trying to avoid the same error.  So I reluctantly call Wisconsin for Biden.

Iowa- Trump won here by 9 pts in 2016. The state’s most reliable pollster (Des Moines Register) shows Trump with a 7 pt lead. Although some polls show a close race I still think Trump wins here.

Minnesota- This state was close in 2016 but the Dems carried it. Biden seems to have a pretty consistent lead here so I think Biden will win it unless the national race is closer than I expect.

THE WEST- Some states (Montana, Wyoming, Idaho, Utah, Alaska) are safe R – others (Colorado, New Mexico, Washington, Oregon, California, Hawaii) are safe D.  The only states worth discussing are Arizona and Nevada.

Nevada- Biden has a small but steady lead here.  Also, Nevada is one of the few states where Dems did better than in the polls. SoI think Biden wins Nevada.

Arizona- Arizona went for Trump by a few pts last year, but today’s polls show it to be close and leaning slightly to Biden (RCP average Biden by 1),,

Voter registration data suggests early voters are fairly evenly divided between Dems and Reps- so that supports calling it for Trump (as does Trump’s 3.5 point lead in 2016).

But… 1) In Arizona the polls were very accurate in 2016 so there is no reason to believe polls are underestimating the Trump vote; 2) the Republicans lost a Senate race they were favored in in 2018 and seem about to lose another one this year, which suggests a trend towards Dems; 3) the Dems are leading in the Senate race there, which could just mean there’s lots of ticket splitting but might also mean some Democrats are willing to admit they are for Kelly (the Dem Senate candidate) but not for Biden; 4) there are 6 real tossup states (i.e. one candidate has a less than 3 pt lead) (NC, Fla, Ga, Tx, Ohio, Iowa and Az) and it is hard to believe Biden will not win one or two of them; 5) the polls favoring Trump (Rasmussen, Trafalgar, Susquehanna) are consistently Republican leaning polls.  So I am inclined to guess Biden wins Arizona.

If I am right, Biden wins 296-242.  But I am very skittish about both my Pa, NC and Az calls- I’d say I have about a 51 percent level of certainty about all 3.  (And also I am assuming that Republican courts don’t upend the results somehow).

By the way, we should have most of North Carolina and Arizona’s results Tuesday except for a few straggling absentee ballots; Pennsylvania will take much longer.  So I think if it is as close as I think it is this could take a while and there is lots of room for court fights etc. 

SENATE
Right now it seems likely based on current polling that the Democrats will lose Alabama, and that (at a minimum) they pick up Colorado, Maine and Arizona.  That brings them to 49 seats.

There are a few tossup seats as well: Georgia, Iowa, North Carolina and Montana.  On the one hand, these all seem like Trump states to me.  But at least two of them (Ga and NC) may be so close that even a tiny bit of split ticket voting may matter.   

In particular, in North Carolina the Democrat, Cal Cunningham, has had a small lead in the last six or seven polls.  So I’m going to go out on a limb and guess he wins. 

Poll data from the other 3 states is more ambiguous, so my guess is that if (as I suspect) Trump wins them, the Republican wins as well.

This means a 50-50 Senate.  (Actually 50-49 since there in Georgia there will be probably a runoff later this year between Raphael Warnock and one of two Republicans; I expect the Republican to win that since Republicans tend to do better in Georgia runoffs than in the first election). 

I note that there are a few other close elections (most notably Minnesota, Michigan, Kansas, and South Carolina) but in those states one candidate (the Democrat in the first two, the Republican in the last two) has had a consistent but small lead and so I think that those leads will hold.

GOVERNORS

The only gubernatorial race that appears to be close is Montana. There it seems like a Republican is slightly favored to pick up a governorship from the Dems.

HOUSE

I don’t expect any significant change in the House, given that the generic ballot test shows Democrats with a significant lead. I’ll guess no change in the House- if either party picks up over half a dozen seats I’d be surprised.

Advertisement

• 1 comment
• Posted under Uncategorized