Princeton Election Consortium

A first draft of electoral history. Since 2004

Jul 13: Biden 380 EV (D+6.4% from toss-up), Senate 53 D, 47 R (D+5.6%), House control D+8.0%
Moneyball states: President AK AR IA, Senate MT KS ME, Legislatures KS TX NC

2020 is a close national election, especially for Senate and state legislatures

May 16th, 2020, 11:37pm by Sam Wang



Despite Trump’s low approval numbers, April/May polls indicate a closer race than Obama 2008 or Obama 2012. This is yet again another close Presidential race, same as it’s generally been since 2000. We live in an era of close elections, and 2020 is not yet an exception. (Note that even Obama’s 53%-46% win in 2008 is fairly close by historical standards. You probably have a different idea of what “close” means.)

A four-point swing – or 2% of voters changing their minds – would make Wisconsin, Michigan, and Florida more likely to go to Trump, and bring the mode outcome close to a perfect toss-up. In the terminology of this site, the current Presidential Meta-Margin is Biden +4%.

A six-point swing (3% of voters changing their minds) would bring the mode outcome to Trump 294, Biden 244.

Presidential poll medians, state by state:
AZ: Biden +9%
PA: Biden +6%
FL: Biden +4%
MI: Biden +4%
WI: Biden +3%
NC: Trump +1%
OH: Trump +3%

Senate. The same is true in Senate races. Thanks to natural gerrymandering (two Dakotas, one California, that kind of thing) and today’s severe polarization, Republicans can maintain control of the Senate with a little less than 50% popular support. Despite that, Democrats have a shot at retaking control if they win a handful of races.

They start from 46 seats (counting Sanders as a Democrat, and assuming Doug Jones loses in Alabama). The basket of competitive Senate races, listed in order from Democratic-leaning to Republican-leaning, is:

Colorado – Arizona – North Carolina – Montana – Maine – Iowa – Georgia (Perdue).

If Democrats were to win four of these races, they would get to 50 seats, opening the possibility of taking control of the chamber. Given current polling data, they could conceivably win all but Georgia and Iowa, which gets them to 52 seats.

In 2020, we will focus on citizen leverage. Your donations and efforts have the most influence on representation in cases where the race is on a knife’s edge – if it’s within a few points in either direction. Therefore Senate races from the middle of the list above are the best place to put resources. In the ActBlue on the right, I currently include Senate races in North Carolina, Montana, Maine, and Iowa.

Conversely, Alabama and Kentucky are not efficient places to put money. I know Democratic readers of this site would like to oust Kentucky Senator Mitch McConnell from office. That is unlikely (at least until new data come in suggesting otherwise), but a more likely way to affect him would be to deprive him of his Majority Leader role.

State races. Finally, I am most excited about our newest expansion effort: identifying leverage in state legislative races and voter initiatives. These are cases where your activism can move the needle on policy and, in many cases, lead to bipartisan rule. for example, in Kansas, changing just one legislative seat would remove the supermajority, and compel the Republican-controlled legislature to work with the Democratic governor. Opportunities to flip partisan control can be found in Arizona, Florida, Minnesota, North Carolina, Pennsylvania, and Texas. These apply to both parties – Republicans would surely like to avoid single-party Democratic rule in Minnesota.
Goal Thermometer
Some of those state legislatures have a say in redistricting. In those cases, a change in control in 2020 would have a decade’s worth of effects. Those states are Florida, Kansas, North Carolina, and Texas. (I’m leaving out Minnesota for reasons I’ll explain another day.)

Again, you can donate to all of these efforts on this site, whether you’re a Democrat (ActBlue) or Republican (RSLC; see right sidebar).

We’ll set up the usual feed soon. Also, the site’s getting a revamp soon!

Tags: 2020 Election

14 Comments so far ↓

  • Dinos Gonatas '84

    it is puzzling that a president as demonstrably incompetent and incoherent as this one has such strong base support (42%), regardless of his policy positions. Any explanations? Psychology of identification with him? How do the demographics break down?

    • Jim Kraft

      Remember Christine O’Donnell (aka “The Witch Lady”) got 40% of the vote in the Delaware Senate race a few years back. I think that 40% is a floor for anyone nowadays.

    • Stanley Chow

      Because they voted for him last time. They don’t want to hear that they voted for a stupid fool, that would made them big suckers. Together with the (exceedingly) low-information voters that they tend to be, and that they are likely at the extreme of the Dunning–Kruger curve; they are going to be rock solid in believing they made a wise choice.

      Incidentally, I like to think this is somewhat like wave function collapsing in Quantum mechanics – if you keep watching a particle, it’s quantum state will never change. On the other hand, if the particle is left alone for a long time, its state can change more easily. So, if they are not reminded constantly that they did something foolish, they will probably forgot the last vote and convince themselves that they in face saw through Trump and did not vote for him.

    • vcujayhawk

      Thank you for reading my mind! Yes, it is absolutely puzzling that any intelligent individual would choose policy positions over competent leadership –especially when American people are dying.

      We are looking at a sitting American President who is decompensating on national television.

  • Dave Daniel

    Okay, I’m following this. I like that you’re being straight about it (and also like that the Dems are looking strong). Thanks.

  • Joseph Barrett Bland

    Thanks for the election update, Sam. And thanks for all your efforts on behalf of protecting Democracy!

    It feels good to see the numbers roll in once more on the site.

  • Jacques

    What about Montana being a Senate pick-up for the Dems? Governor Steve Bullock leads by 7 points in the latest polls. He’s won 3 consecutive state-wide races (Attorney General and Governor x 2). Much better pick-up chance there than in N.C., Georgia, or Kansas.

  • Abe Fisher

    We live in an era of “close elections” because the Electoral College has diverged from the country at large. Because we haven’t increased the size of the House of Representatives in about 100 years, larger, typically bluer states are underrepresented, while smaller, sparsely populated, typically redder states are overrepresented. If we were expand the House so that the state with the smallest population got one Rep., and then that state’s population was the target size for all congressional districts, the popular vote would continue to be as close-ish as it has been, but the elections themselves would not be.

  • Karl Chwe

    Nice to see the focus turn to November! I can’t believe I’m saying this, but the election may be more important than 100,000 COVID-19 deaths (guaranteed) or even 200,000 (likely.)

    “The basket of competitive Senate races includes Arizona, Colorado, Georgia, Iowa, Maine, Montana, and North Carolina.” I.e., seven states.

    “Therefore the six Senate races above are the best place to put resources, as listed in the ActBlue on the right.” Six? (The ActBlue page mentions only five, and only three senate races.)

    Typo?

    Thanks for everything you do, Dr. Wang!

  • Marco Ciocca

    In 2016 Hillary had a meta-margin of roughly +2 at the end (if memory serve). Trump won enough of the “I dislike both” to break through in PA, MI and WI. I wonder if this time the dislike both camp will break more for Biden. That would probably make the current meta-margin of +4 more like +5 (or 6). Clearly this is punditry and not statistical analysis but one can hope :)

    • Sam Wang

      The final projection in 2016 was a Meta-Margin for Hillary Clinton of +1.0%, i.e. the point estimate was 1% above threshold for 270 EV. So the final error was somewhat less than 2 percentage points of Meta-Margin.

      http://election.princeton.edu/2016/11/08/final-mode-projections-clinton-323-ev-51-di-senate-seats-gop-house/

      Yes, there is survey evidence that voters who dislike both Biden and Trump will break toward Biden. But one should be careful about spending that money before one had it! For one thing, we don’t know if they behave the same as “undecided” voters (in quote marks because self-reported undecided voters are probably predictable, they just aren’t aware of their own preference yet).

  • Will Hutchinson

    First, please remember to count Angus King as caucusing with the Democrats, too.
    Second, all but Iowa and Georgia gets the Democrats to 51, assuming Jones loses, not 52.
    Third, there has been a lot of news in the past month — Jon Ossoff winning the Democratic primary in Georgia outright, the Des Moines Register’s recent poll showing Greenfield leading Joni Ernst (albeit within the margin of error), and Kelly Loeffler continuing to take incoming fire over her stock sales. Do any of these or other events change your priorities for seats likely to flip?

Leave a Comment