It has long been an aphorism that presidential elections don’t start until Labor Day. And in some sense, it is still true. The summer conventions draw fewer and fewer viewers every cycle, and this year’s stripped-down extravaganzas will likely be short on compelling theater. Donald Trump has threatened to defile Gettysburg with his nomination acceptance speech, though it remains to be seen whether that threat was only to make his preferred option, the White House itself, more palatable. Once defiled, always defiled, but perhaps one hallowed land can survive this Administration.
In the meantime, it may be helpful to explain the state of the 2020 election map. First, though, the 2016 map:
When I left the office on Election Day 2016 for the train ride home, the Clinton campaign was reporting optimism in Florida. By the time, the train was in El Monte, it was clear things were going wrong in Florida. And by the time the train stopped in Claremont, it was clear Pennsylvania would put Trump over the top. For the first time, uh, ever, I skipped election night and took the kids to see Doctor Strange. By midnight, it was apparent that Wisconsin and, unbelievably, Michigan had followed Pennsylvania, while New Hampshire was on a knife’s edge.1Trump ended up losing NH by about 3,000 votes.
None of these three states Trump carried had voted Republican since 1988. Meanwhile, Trump carried Iowa–historically, a swing state–by nearly ten points. The same was true in Ohio–a more reliably Republican state than any of the others, but one Obama won twice. To recount all the reasons–both real and imagined–for these results is not the purpose here. Suffice it to say that the Great Lakes states have a high share of non-college educated whites, and minority voters (or more accurately, non-voters, particularly in Michigan) ended up as unenthused with Clinton as most others.2Yes, many, many reasons, several ignoble. We’ll discuss elsewhere.
Nevertheless, Trump’s victory was statistically unlikely. He won Michigan by 19,000 votes (.23%), Pennsylvania by 44,000 votes (.72%), and Wisconsin by 22,000 votes (.77%). It is not hyperbole to say that he won by a perfectly distributed 85,000 votes out of over 136,000,000 cast. On top of that, he won Florida by 1.2%, and Arizona and North Carolina each by 3.5%.
It is certainly possible that Trump can pull another inside straight. The combination of non-college white males and the electoral college is relatively potent, hence Trump’s recent emphasis on the Confederacy and college football. But Biden remains favored as the campaign nears the home stretch.
The easiest path to Biden’s victory is shown above, simply recapturing the historically Democratic Great Lakes states. Polls in each of these states give Biden a statistically significant lead, and regression to the mean already favors him.
Before proceeding further though, let’s pause and consider the number of variables that produce a presidential election. First, 200 million registered voters must decide whether they will vote at all. Then, they must decide whom to vote for. There are the inevitable blizzards, snowstorms, traffic jams, lost ballots, voting booth lines, and the dozens of other imperceptible events eventually aggregated and reflected, however imprecisely, in the final count.
There is no way to control the risks inherent in such a variable exercise, but their effects can be diluted through diversification. In February, with Trump hitting peak popularity, it looked as though Biden would have to sweep the Great Lakes states while holding the states Clinton won in 2016. For example, one could have imagined a map where Biden recaptured Michigan and Pennsylvania, but lost New Hampshire and Minnesota–two 2016 Trump near-misses–and held Wisconsin. The result would be a narrow Trump electoral vote win.3Given that Trump is looking at 25% in California, there are no scenarios for him to win the popular vote, short of actively interfering with voters, and when would he do that? In short, Biden needed to do only a few things to win in that scenario—but he had to do them perfectly.
But such a scenario is now becoming unlikelier by the day. The 2020 election instead looks to be fought in a series of 2016 Trump states, expanding Biden’s potential victory scenarios.
For example, the scenario above demonstrates a potential result should Biden capture Michigan (as is likely), but lose Pennsylvania and Wisconsin. Biden can make up that deficit by winning Florida (alone), or in tandem with Arizona, both states where Biden leads the polls. The result is a narrow but safe electoral college victory.
But the factors placing Arizona and Florida in play are also present in other southern states with large urban centers and diversifying populations. In other words, North Carolina and Georgia are right behind. If Biden adds these states and tacks on Pennsylvania and Wisconsin, the resulting 350 electoral vote victory would be decisive. Further, this scenario almost certainly elects a Democratic Senate, as Democrats flip Arizona, Maine, and North Carolina–with two toss-up Georgia Senate seats to contest as well.
And then there is the potential for generational realignment. Unlikelier still, but possible.
In this scenario, Ohio and Iowa fall with the Great Lakes states and Texas makes its stunning debut as a two-party state. In the fallout, Biden wins Louisiana in a recount. Republicans also show weakness in Utah, Kansas, Missouri, and South Carolina–four states with state GOP infrastructures far more conservative than the emerging suburban consensus is willing to go.4Trump’s cartoonish bullying of “suburban housewives” notwithstanding.
There are also nightmare scenarios, of which we must consider at least two reasonably possible ones, if only because 2020 will show us no mercy.
These scenarios require a bit of explanation. Maine and Nebraska award some of their electoral votes by congressional district. Each state’s overall winner receives two electoral votes, while the winner in each congressional district wins an additional electoral vote. This has not been historically relevant, but in 2016 Trump carried ME-2 while losing the state, and won NE-2 (in effect, the city of Omaha) by only 2.5%. If Biden wins Pennsylvania, Michigan, and NE-2, and all else remains as is, a 269-269 tie results. The same is true in the less likelier scenario that Biden wins Pennsylvania, Michigan, and Arizona, but Trump wins Minnesota. There is a constitutional procedure for resolving electoral college ties, but it will be hard to execute in the midst of mass ritual suicides.
What does this all mean? None of it is predictive, but it is illustrative. Joe Biden must win 38 more electoral votes than Hillary Clinton to take the White House. The more electoral votes are in play, the more likely he is to obtain them. In other words, a lot of uncertainty in the weeks leading up to Election Day may suit Democrats just fine.