Last week, The New York Times and Siena College published polling that showed former Vice-President Joe Biden ahead of President Trump nationally 50–36 (the same margin the president went to war with in an earlier CNN poll). More critically, Mr. Biden led Mr. Trump by double-digits in Michigan, Pennsylvania, and Wisconsin — all states the president carried four years ago. Mr. Trump’s responses to the coronavirus outbreak and the Black Lives Matter protests have dogged him since at least the end of May, strengthening Mr. Biden’s lead over him polls both across the country and in 2016’s battleground states.
When a candidate leads by over ten, fifteen points in any race, the race hardly seems competitive anymore. This is especially true when it’s multiple high-quality polls showing roughly the same result. Colorado, for instance, is favouring Biden by 17.7 points according to FiveThirtyEight’s polling average. Similarly, Mr. Biden is outpacing Mr. Trump by an average of 10.7 points in Michigan, including the 11-point lead from the Times/Siena poll. These absolute leads fall so far from their polls’ margins of error that multiple top-notch polls would have to be wildly off for these states, which were close in 2016, to actually be close again this year, let alone for Mr. Biden to lose them.
Yet these states remain “battleground” states to many commentators. A recent FiveThirtyEight piece included Colorado, Maine, and New Mexico on a list comparing “battleground states”, despite their fourteen-plus Democratic leans. CNN’s Electoral College map rates, as of 1 July, Maine and New Mexico as safe blue states, yet Colorado and Virginia only lean blue here (even though Virginia is another state where Mr. Biden is, on average, ahead by over ten points). NBCNews puts Michigan in toss-up territory with Arizona, despite the latter being a much closer race. 270toWin is reluctant to say Florida, Wisconsin, or Pennsylvania even lean one way. Same for the Cook Political Report.
It’s safe to say that outlets are choosing to view this race with 20/16 vision. The shock of Mr. Trump’s upset victory is still reverberating throughout the mediasphere, leading to an over-correction this time around in how states are projected to account for the mistaken predictions four years ago. I won’t get into why mainstream interpretations of 2016 are off here. However, considering that Democrats went on to win back the House in 2018 and are making the upper chamber competitive this year, the political weathervanes point to a national environment that trends blue in 2020 (with generic ballot polls confirming that).
What’s the harm in being conservative in state predictions, though? Don’t you want to err on the side of caution when it comes to something as potentially volatile as a presidential race? More than that, many of these states getting labelled as swing-y this year have been swing-y in the past. Would you want to suggest, for example, that Michigan going for Mr. Biden is somehow assured, or even inevitable, when it went for Mr. Trump last time? This wouldn’t be about living in 2016’s shadow: rather, it would be a reflection of the historic trends that takes into account that we are more than four months from Election Day.
The problem with this thinking is that it simply does not reflect our current reality. Let’s look at two states: Colorado and Kansas. Colorado is considered a toss-up in some of the electoral match-ups I cited above, despite that polling average of +17.7 for Mr. Biden. Kansas, on the other hand, averages at just +8.8 for Mr. Trump. I challenge you to find the electoral map that considers Kansas a toss-up at this juncture, though. Senate races this year in both states also reflect these blue-ing trends which do not bode well for Mr. Trump in his race. No, Kansas has not voted Democratic at the presidential level in decades, nor do I see it doing so this year. Rather, what I hope to show here is that the reason you can have Mr. Biden up in Colorado by a higher margin than Mr. Trump is up in Kansas, yet consider the former a toss-up and the latter solidly Republican, is out of the optics of 2016 (I’ll add that Kansas is actually more elastic, or movable electorally, than Virginia).
The other issue to combat here is the thinking that posits any trends or data we have at this point should not govern our conjectures for November. Of course this isn’t November. We have no idea what exactly will happen over the next few months. For as crazy as this year has already been, Martians could land and try to invade New Jersey in October and I wouldn’t be surprised. Remember the Iran controversy from January that sparked threats of war? How ‘bout that impeachment? The national scene could shift completely in the ensuing months for how unpredictable 2020 has been so far.
The caveat here is that, while there will certainly be new news events and potential crises that arise, the fundamental priors will likely hold. CNN was reporting back in May how Mr. Biden’s lead in national polling is the “steadiest on record” — in absolute terms, that is still true, even with his advantage ticking up since then. Given how he led throughout 2019 in the Democratic primary and held that lead, with some brief interruptions, to win the party’s nomination, we shouldn’t discount the former vice-president’s ability to maintain ground once he has it. As far as Mr. Trump goes, public opinion of him has been remarkably steady, moving little since February 2018 (besides when the government shut down in early 2019 and, of course, now). In addition, a YouGov poll published on the 26th shows just how deep Americans’ views of both politicians lie: when asked, “Would you say that your mind is made up about how you will vote?” between the two candidates, 94% of respondents agreed. When voters’ impressions are set this far out from November, you should feel safe making some assumptions about how they will act. Let’s not forget as well that we are approaching a depression, which aren’t typically good for incumbents.
More importantly, some aspects of this race are just not going to change. Given the nature of the novel coronavirus and the current administration’s continued fumbling response to the outbreak, COVID-19 will remain part of the political calculus well past November. This will stay a sticking point for the president, then, since he has not received good marks on how he has handled the situation through now and will likely do nothing to change that trajectory. Indeed, the greatest constant in politics today is Mr. Trump himself, who has shown us time and time again, despite initial insistences to the contrary, that he prefers playing to his small base than backing down from positions and actions that are unpopular.
So what can we say about the present, if we acknowledge that Colorado and Michigan today stand out of Mr. Trump’s reach? What are the swing states in 2020? If even Minnesota is out of the mix right now, where do we turn our attention? This poll from Talk Business and Politics-Hendrix College released in the first half of June gave me pause when I first came across it, and was the origin of this question. It found a majority of residents in Arkansas disapproving of the president’s job 50–46 and voting for Mr. Trump by just two points over Mr. Biden if the election were held today. That’s a two-point difference in a state the president carried by almost twenty-seven points four years ago. Granted, this was one poll in a state that has received little attention otherwise; still, for all the talk about Georgia and Texas being in play this time around, the sheer possibility that Arkansas might be this close is an indication of how the landscape has changed for Mr. Trump from 2016.
With this information in mind, I compiled a list of states I am keeping an eye on through November. These are states that have had enough polling done for FiveThirtyEight to create polling averages for them, with those averages indicating close races of ten points or less. They include, as of 1 July at 2:00 EDT:
- Nevada (+8.7 for Biden | -0.9 relative to nation* | 2016 margin: 1.5% [Democratic win])
- New Hampshire (+8.4 | -1.2 | 0.4% [D])
- Wisconsin (+8.1 | -1.5 | 0.8% [R])
- Pennsylvania (+8.0 | -1.6 | 0.7% [R])
- Florida (+7.3 | -2.3 | 1.2% [R])
- Arizona (+4.8 | -4.8 | 3.5% [R])
- North Carolina (+2.9 | -6.7 | 3.7% [R])
- Ohio (+2.7 | -6.9 | 8.1 [R])
- Georgia (+1.7 | -7.9 | 5.1% [R])
- Iowa (+0.1 | -9.5 | 9.4% [R])
- Texas (+0.2 for Trump | -9.8 relative to nation | 2016 margin: 9.0% [R])
- Missouri (+2.4 | 12.0 | 18.6% [R])
- South Carolina (+7.3 | 16.9 | 14.3% [R])
- Utah (+7.7 | 17.3 | 18.1% [R])
- Kansas (+8.8 | 18.4 | 20.6% [R])
*This number is calculated by taking the state’s polling average and comparing that to the national average, which is +9.6 for Biden.
Of course, the map becomes a little pointless to follow once you put this collection of states up for grabs. Mr. Trump could only afford to lose a few low-hanging fruits before the numbers reached 270 for his opponent, not to mention having to hold onto all the Republican-leaning states, which he carried last time. On the other hand, if Mr. Biden wins all the Democratic-leaning states here along with the states already polling more than ten points for him, he would fast approach a surplus of one-hundred electoral votes; add the red states, and the vote nears 450. This would become a landslide the size of which the country has not seen since the 1980s and indicate an astonishing reversal of 2016.
Do I think a landslide of these proportions is likely? Heavens no. But within the realm of possibilities, this scenario is more likely than Mr. Trump winning all the states I have listed, if only because of his downward trajectory since May. CNN has similarly begun to opine on the chance of a landslide, albeit of smaller proportions, with this article being the first I’ve seen to include Missouri as a possible flip. Sure, it’s insane, and you would think we would have become too polarised as a nation to vote in such a uniform fashion anymore, yet in a year where Texas is already polling within the margin of error, we might just see another Reagan-era “swimming pool”.
So, my advice for those watching the presidential race right now: keep adjusting your priors. And my advice for pollsters: poll Arkansas more, not New York!