The COVID Realignment

How our Political Parties Became Reshaped along an Axis of Risk Aversion

A key focus of this newsletter has been to think about the ways that Covid has reshaped our lives in ways that diverge based on occupational status. We’ve talked about how the Zoomer WFH class has fled cities, while essential workers have stayed behind to work and therefore been exposed to greater disease risk. Today, I’m going to argue these divisions are now also creeping into our politics.

The Populist Revolt

First, we need to recap how 2016 realigned our politics. As we all know, Trump really changed the base of Republican support towards working class whites in the Midwest. Neither anti-immigration nor anti-trade appeals had generated a lot of traction in these groups in the past. But, for whatever reason, combing these two things into a broader meta-message of “borders” did. So, on net, losing a few Greenwich Republicans was a very good deal for the Republican party as it built a more populist support base which was very optimized for the Electoral College.

It was unclear if this was a deeper shift in our two political parties which would persist past that particular cycle. It could be that Hillary Clinton was just caught in the middle in 2016, when Democrats had lost their former union labor strongholds, but had not yet developed a base in the growing Sunbelt regions to push through the electoral map. Key to the future of populist politics is what happens with non-white voters. Would the Republicans be able to build a durable cross-racial support base among the working class, or would their reliance on white voters doom them to irrelevance with changing demographics?

Growing Appeal for Republicans Among Non-Whites

So I started looking at some demographic crosstabs of voter intention in 2020 to try to figure this out. An important caveat here is that we are not always able to line up precise surveys across years, so it’s hard to tell which trends are really robust. And of course you don’t know how intentions today will line up with actual voting behavior. But a fairly strong trend is — Trump is picking up substantially more non-white support in 2020 than he did in 2016. In fact, by some measures this will wind up being the least racially polarized election in at least 20 years:

My initial thought was the racial patterns might follow the broader polarization in our society around education, geography, and so forth. So perhaps Trump was starting to see gains among Black and Hispanic voters with less educational attainment, the same way that he did starting with white voters.

This turns out to be wrong — growth in Republican support is strong among College educated non-whites. From 538:

In fact, even among whites — the more educated are turning towards Trump!

So it’s not just racial polarization, even education polarization which was a really big deal (Chris Arnade, “Back Row Kids”) has kind of gone away.

I think these underlying political shifts are actually pretty crazy if you think about them. We’ve had four years of… let’s say fairly racially charged rhetoric from our political parties. Years of outcry against our latte-sipping liberal elites, the deep state, fake news, etc. And yet people’s actual voting decisions are less connected to their identity in terms of their racial or educational background than before. But it really doesn’t seem like partisanship is down at all. So what is driving polarization today, if not the standard forms of identity?

A Major Political Axis is now Coronavirus

I argue that it’s increasingly the coronavirus itself that’s driving our political decisions. On one hand this may seem fairly obvious — this is clearly the biggest thing going on right now, so it’s natural for it to feed into the political process. However, the conventional wisdom is actually closer to the idea that Covid hasn’t really changed that much in the race. For instance, people point to the stability of Biden’s lead, which dates back to before the pandemic:

My argument is that, behind this apparent stability, there is actually a lot of churn which has sort of netted out in ways that have left our electorate less polarized on existing identity dimensions, but more polarized on a new (and harder to observe) dimension which is something like “risk tolerance to Covid.”

What’s the evidence for this? First, let’s look at how different populations perceive our Coronavirus response and how we should think about “lockdowns” vs. opening up.

In terms of sending children to school, postgraduates now cluster with Republicans in wanting their children to go to school:

This results in slightly odd politics here. Higher-income and higher-educated workers really want their children to be in school, and often motivate school opening by arguing that it’s essential to avoid educational disparities. But lower-income and less-educated workers themselves are actually much more risk-averse themselves when it comes to school reopening; plausibly because the disease has just run through their communities.

The same is true across a range of beliefs both about how bad Covid is; how people behave themselves; and the role of policies combat the disease:

People engaged in essential work and otherwise at risk of contracting Covid are much more worried about disease risk and want us to take much more action on public health. Some people here are at least somewhat insulated from economic costs given social security support, and the other range of support we did in the CARES Act. Still, I think it’s probably the case that these groups are probably willing to sacrifice a certain degree of economic well-being for better health.

Then, we have other workers — who disproportionately work from home, or are managers and so forth — who really want to open up the economy. Some of these people plan to start going out and about afterwards. But they may also really want their stores to start seeing more customers and so forth. They want schools open so their kids can get out of their home offices. And they have health insurance coverage through their employers to support them. This lines up with the gender divide too, which has remained pretty large; since women on average think that the pandemic is more serious than men.

A New Resorting

This shift leaves many people who are now misaligned in their party based on their personal risk tolerance and views on Covid. I think these people then start to shift political affiliation. So among higher education Blacks, whites, and Hispanics — we see greater shifts to Democrats. Whereas, the elderly and less educated whites are voting for Biden.

One way of looking for this is checking what issues motivate voters. Republicans retain a large lead on economic issues, while Democrats are preferred on health and the coronavirus. So voters who really prioritize the economy as an issue, and think opening is important, will gravitate over time to Republicans; and similarly voters concerned about Covid risk will become Democrats.

This popular conception, of course, is at odds with the typical economist view, which is something like “We have to deal with the pandemic to open up safely.” It doesn’t seem like most voters perceive it that way, and instead see a large tradeoff between opening up and accepting more Covid deaths, or staying more closed to prioritize health benefits.

It’s also worth noting the particular partisan divide we have here wasn’t set in stone. Back in the early Spring — Democrats were arguing that the virus wasn’t a big deal, and were arguing against travel restrictions. Trump was actually arguing in favor of masks at one point, and Tucker Carlson supported them too — back when the experts were against them. That’s all completely flipped now.

What Does This Mean For the Election?

If I had had to guess, I would say that the reshaping of parties into one branded as “personal safety” and another as “opening up” is good for Democrats. It reverses some of the shifts in the last election, and so gives Democrats a further edge in Midwest states with large numbers of older and white working class voters. At the same time, it helps Republicans solidify leads in places like Texas and Florida, where they are seeing declines in support from older voters offset by gains among Hispanic voters.

These trends are amplified by where the pandemic is strongest right now. In the Upper Midwest, Biden is seeing even stronger polling — where in places like Florida where the disease is receding, Trump seems to be narrowing the polls.

I’ll admit I don’t have conclusive evidence for the “coronavirus realignment” hypothesis — I would like to see what issues party-switchers prioritize. I figured I would write this post to pre-register the hypothesis, and revisit after the election once Exit Polls are out. Please also send along any data which might speak to the idea one way or the other.