Congratulations in advance to the Democratic nominee for the presidency in 2020, who, The Fix can now project, will likely win election (or reelection). That projection comes with two big caveats: 1) States have to keep voting exactly the way they have been in recent history, and 2) the national popular vote is dead even.
Last month we analyzed an interesting point of data: the extent to which states differ from the rest of the country in their picks for the presidency. For example, take Indiana in 2008. The state went for Obama by about 1 percentage point while the country backed Obama by 7 percentage points -- making Indiana six points more Republican than the rest of the country. By tracking the relationship between a state and the rest of the country over time, you can track how its politics are changing relative to the nation as a whole. In that July post, we noted that Virginia has moved quickly away from more strongly supporting Republicans in recent years, but that Georgia has become less Republican than the rest of the country at a much slower pace. The argument that the South was turning rapidly toward Democrats, we reasoned, was a bit premature.
We decided to take that thought experiment a step further. For each state, we looked at 100 years of its presidential voting patterns, comparing the margin it supported the Democratic or Republican candidate with the the margin that candidate had in the rest of the country. Then, for fun, we extrapolated out the last four elections to try and gauge where each state might be in 2020.
Before we get too far, we'll acknowledge that this is an extremely rough guide. It doesn't account for any sudden shifts in demographics; it doesn't account for things like candidates doing better in their home states. We ran a similar extrapolation using data from 1992 to 2004 to see how the results matched up with 2012. As you can see at right, there's a strong correlation between what we would have expected and what resulted, meaning that we can at least have some confidence in our 2020 projections. (One outlier is labeled on that graph: Hawaii, which outperformed expectations, thanks to native son Barack Obama on the ballot.)
In each of the charts below, the vertical axis goes from a state being far more Democratic than the national average (at the top) to same as the national average (zero) to far more Republican (at the bottom). The horizontal axis shows the year of the election -- with 2020 the dot at far right. So in the case of Alabama, below, it used to vote more Democratic than the rest of the country, until 1984. Now it's more Republican (hence the red line), and by 2020 it would be over 36 percentage points redder than the rest of the country if the trend holds.
So. Using those 2020 extrapolations, we created four categories.
The big movers
States that will grow significantly more Democratic or Republican.
Here's one place that the extrapolation will raise an eyebrow: Hawaii is projected to outperform the rest of the country by 58 percentage points for the Democratic candidate. That's the result of projecting Obama-level enthusiasm out for eight more years. That's not likely to happen.
One that shouldn't be a big surprise is West Virginia. While the extrapolation probably overestimates the extent to which the state will move right, it seems unlikely it will completely rebound to the Democrats' benefit.
The swing states
Any state that was projected to be 10 points or fewer from the national average in 2020 falls into this category. These are the states that will be closer than the others, in other words.
A lot of these "swing states" won't surprise you: Florida, Ohio, etc. But some are interesting. Montana is probably closer than it should be, but the idea of Democratic-tilting Missoula taking over the state's politics is interesting. Also worth noting: We're using a pretty flexible definition of "swing" here. Montana would still be six points more Republican than the rest of the country in this projection.
Notice, too, how many of the closest states are "blue" ones -- that is, more likely to support the Democratic candidate than the national average. That will come up again below.
The solid reds
Any state that is 10 points more favorable for Republicans or more.
There's Georgia again -- less Republican than many of the solid red states, but not totally. And there's Texas. A reminder: This extrapolation doesn't include any demographic considerations. So if, say, Latinos register and vote Democratic in Texas more rapidly than they have in the past, that little dot on the right will shift upward more quickly.
The solid blues
Any state that is 10 points more favorable for Democrats or more.
Notice that, in this formulation, Nevada's evolution into a blue state continues. In part, that's thanks to demographic changes that have been incorporated into past elections. (See also: Virginia.) The extrapolation moves Colorado into solidly blue territory, continuing its recent trend. We shall see.
So now we come to the contentious part: the predictions. We know the electoral vote totals for each state in 2020, since they won't be reallocated until after that year's Census. If we assume, then, a dead even vote, 50 percent of the country backing the Republican and 50 percent backing the Democrat, we can say that the Democrat will win reelection, with 300 electoral votes. After all, every state that is more Democratic than the national average will then have gone for the Democrat, and vice versa.
But that changes quickly, depending on the overall margin of victory. If the Republican candidate does 0.4 percentage points better on the national average, suddenly the Democrat only leads by 25 electoral votes. If the Republican margin goes to 0.5 percent, he or she wins.
There's a much bigger caveat: A small variance in these projections makes a huge difference. Ohio and Pennsylvania are only barely more likely to be more Democratic than Republican. If the actual vote varies from that -- which it will! -- the numbers shift a lot. And again, more of the closer states are in the Democratic category than the Republican one.
What we're saying is this: If you bet on these results, please do not knock on our door on November 4, 2020, demanding your money back. But if these predictions are totally correct, we do expect a cut of the winnings.