Political and social correlates of Covid-19 mortality

Do political and social features of states help explain the evolving distribution of reported Covid-19 deaths? We identify national-level political and social characteristics that past research suggests may help explain variation in a society’s ability to respond to adverse shocks. We highlight four sets of arguments—focusing on (1) state capacity, (2) political institutions, (3) political priorities, and (4) social structures—and report on their evolving association with cumulative Covid-19 deaths. After accounting for a simple set of Lasso-chosen controls, we find that measures of government effectiveness, interpersonal and institutional trust, bureaucratic corruption and ethnic fragmentation are currently associated in theory-consistent directions. We do not, however, find associations between deaths and many other political and social variables that have received attention in public discussions, such as populist governments or women-led governments. Currently, the results suggest that state capacity is more important for explaining Covid-19 mortality than government responsiveness, with potential implications for how the disease progresses in high-income versus low-income countries. These patterns may change over time with the evolution of the pandemic, however. A dashboard with daily updates, extensions, and code is provided at https://wzb-ipi.github.io/corona/

That is from a new paper by Constantin Manuel Bosancianu, et.al., via Alex Scacco.


The paper is trying to say something, but what? Gobblygook.

Bonus trivia: compare Israel (second wave), Chile (exponential! Did they learn nothing from the Chicago boys?), Mexico/Brazil and India/Pakistan with Greece, New Zealand, South Korea.

Its excluding hypotheses.

My evolving hypothesis is the rate of spread is driven by global and regional social and business mobility first, then by mostly forced trade contact, followed by local social mobility.

Ie, Seattle, Boston, and NYC were seeded early by business and social global travelers, who then spread it to their mobile peers regionally, who then spread to the workers forced to work in contact with them.

The spread through the workforce has been slow, but once it enters into the heartland areas with concentrations of high contact workers, eg meat packing plants, local social contact spreads in through the community.

The shutdown in coastal areas cascaded through the US so even if the 40 States not infected didn't lock down, the 10 States that did depressed 60% of US production hitting the other 40 States hard.

Now as the coastal States have some control over the epidemic, the other 40 States are being hit with virus by they are restarting from the external imposed shutdown, so a local lockdown to deal with widespread infections and overload of hospitals are politically impossible.

It will take six months to see how close to reality my hypothesis is, especially as deaths are not reported quickly and correctly attributed to COVID-19. Where conservatives have shutdown hospitals, the deaths will be at home, and only attributed to COVID-19 is that will bring in Federal cash to local governments who are willing to accept the strings and political stigma.

Page 18 seems both legit and sad.

And it seems to repeat the one-two punch of our reality, that populism undermines institutions, but you can't complain about that, because "that's partisan."

Populism rises as institutions fail.

That's not what page 18 says.

Populism is stupid. That’s not particularly partisan.

For those of us who made it past the literature review of political science in the paper, unlike anonymous here, it’s also not particularly relevant

Based on the results of the paper populism happens to be both stupid and orthogonal to Covid outcomes.

Which is entirely unsurprising given what we know from public choice

I don't think it's a very compelling argument that we should turn off the news, and forgot everything we actually know about the US response to COVID.

Especially not for the tautology that any response, perhaps every possible response, was just "public choice" anyway.

A theory without any predictive power, except in the rearview mirror.

Case in point,

U.S. axed CDC expert job in China months before coronavirus outbreak

That job was clearly axed because in-group and out-group drama, and not because Public Choice got involved.

Unfortunately, as per usual everything you’ve said is a combination of incorrect statements and intentional lies. Let’s review your bullsh*t:

a) I never said to turn off the news, this is an irrelevant strawman

b) we shouldn’t forget the US response to pandemics, which is why I would never say that. It’s a perfect lesson on Public Choice in low trust societies

c) Public Choice isn’t a tautology, you don’t seem to understand what that word means

d) Public Choice has extreme predictive power, it’s the one discipline that’s been 100% correct in predicting everything from the TSA, police violence, the Iraq war, the success/failure in Covid response by country, etc. It was obvious the US would fail. It’s been a dysfunctional civil society for decades and its institutions are a rotting corpse of corruption, bloat, mission creep, and extreme partisanship

e) the CDC link you gave is weak, but you’re unintentionally giving further examples of Public Choice in action, so lol

Sigh, Boomer

Time to pack it up, and go to a country that really understands how public choice is not a tautology but instead is the one discipline that has allowed it to become superior to the failing United States.

Me: American Colonies

"That was public choice!"

Me: American Revolution

"Public choice!"

Me: Slavery

"Public Choice too.."

Me: Civil War

"ah, Public Choice"

Me: Emancipation

(mumbles) "public choice"

Me: Jim Crow

"Public Choice!"

Me: Civil Rights

".. communism!"

Me: Populism

"Public Choice!"


"More of that communism coming to destroy our way of life!"

anonymous once again thinks that creating a strawman position is a valid argument.

But broad based support for me. Good luck defining the difference in terms everyone will accept as being non-partisan.

anonymous is highly partisan, he just can't comprehend how his own affiliations are partisan. In his mind, his preferences are clearly morally correct and the right choices and anyone who disagrees is clearly of malicious intent and/or dumb. Ergo, he's not partisan, but everyone who disagrees with him is.

I quoted the definition below. It sure as heck ain't another name for democracy, and the principle that all men are created equal, all citizens equally valued.

If Democrats run your state, you were....

Populism doesn't wreck institutions; it exposes them. The populist uprising of the last couple of weeks did so for our epidemiological and medical authorities, revealing them to be nakedly political in what activities they choose to permit or prohibit. A wealth of scientific credibility lies in ruins not because of populism but because the people who run our institutions threw down their professional reputations in order to chase transitory Twitter clout.

This sounds like Anthony Fauci's and the Trump Administration's sudden clinging to the IHME model in early April. In contrast to other models, this one predicted only 60,000 deaths this year from covid-19 and became obsolete just a few weeks after Fauci trumpeted it.

Why did he do that, except to provide political cover for Donald Trump?

heres a good example of elite leftist public health stupidity
deblasio is hiring "contact tracers" & prohibiting "contact tracers" from asking if covid 19 patients attended demonstrations, thus fubaring the
definition of contact tracing

If Manuel were Maria; there's no place with stronger male egos than in economics: https://www.nytimes.com/2020/06/13/opinion/sunday/women-leaders-coronavirus.html

Kristof's article is shockingly stupid even for the NY Times.

After a quick look at the paper and a few thoughts of my own. Governments in countries with:

Prior experience in controlling SARS acted rapidly to respond to covid19

Societies with high social capital followed expert advice (both right and wrong in hindsight).

Countries with high state capacity responded quickly with test-trace-isolate.

Countries who doubted their state capacity to respond locked down quickly.

No nonsense about female leadership being good and populist government being bad.


In a definition given and operationalized by Kyle and Meyer (2020), populist political leaders are united by two claims: “(1) that a country’s ‘true people’ are locked into a moral conflict with ‘outsiders’ and (2) that nothing should constrain the will of the ‘true people’.” Under this account, populist politicians can exacerbate cultural divisions, take a skeptical position towards science and expertise, and can be erratic in decision-making, eschewing the policy moderation that can result from between-party compromise. Insofar as populist leaders are motivated by electoral considerations to take anti-establishment positions, they can also weaken state services, such as healthcare provision, that rely on institutionalized bureaucratic structures and expertise.

Have we really gotten to the point where we have to say "this is fine" in order to be .. what, objective?

The current politicians contributed to a given country's response to covid19, but the paper is saying there stance (whether populist or not) seemed pretty irrelevant.

Are arguing a case that populism in the past has lead to low current social capital and/or state capacity? The paper does indeed associate these with poor covid19 progressions.

I meant " Are you arguing a case . . . ."

This strikes me as an "ignore the blood on the floor" argument.

There is plenty of awfulness to be seen. If this model does not capture it .. well, did we actually whistle past the graveyard with COVID-19, or is the measure .. incomplete?

You’re quoting the paragraph where they describe the *hypothesis* that populism has pernicious effects on response, and background on said hypothesis. In their results they find that populism did not have an effect on response.

Stop quoting the hypothesis as if it were the result.

Correct observation.

There is actually a difference between a hypothetical and an observation.

anonymous doesn’t understand how academic papers work.

He believes the literature review is the result.

And it becomes a perverse defense of awfulness to insist that observation is just hypothetical.

The hypothesis doesn't really exactly seem too likely to me.

Populist politicians *could* take a more skeptical stance to scientists, but there is simply no relationship at the mass level with "trust in science" and "populist" governments. Certainly many "populist" politicians would have been more willing to close borders, earlier, with free choice.

As for "exacerbating social divisions", well, social divisions exacerbated pretty well before on their own under conditions of heavily pro-minority, pro-bureaucrat government.

Geez, what was I just saying about the difference between a hypothetical and an observation?

Trust in Medical Scientists Has Grown in U.S., but Mainly Among Democrats


"Public confidence in scientists has remained stable for decades".

No change 2016-2018 for sure.

Some change, I'm sure, among the left winger LARPers who think the lockdown had a scientific basis and signal wildly about how much they love science, but otherwise no change.

I think you chose a poll which did not look at the political dimension. Perhaps "public confidence" is the same, but "sorting" is different?

Democrats and Republicans differ over role and value of scientists in policy debates

See below; party sorting changes direction, but by modest amounts, with Repubs where Dems were in early 90s (and what a disaster that was for science policy?).

Re the other article, I'd suggest that the responses of epi applying knowledge in different circs indeed vindicates the view that "scientists are just as likely to be biased as other people".

For political affiliation and change over time - https://phys.org/news/2019-11-americans-high-science.amp

No decline effect and very slight change, even by party.

There is a very clear tick there, 2016-2018. And it is interesting to note the current ranking. Currently Democrats are ~ 51% on "a great deal of confidence" in science. Independents sit at ~ 42%, and Republicans at ~ 38%.

So stop fooling around. The populism has created a shift, or perhaps more likely a new sorting. The 38 to 51 percent difference is significant.

The shift is all within the Democrats and "towards" science.

So it's in the opposite predicted direction of "populism" causing a decline.

lol, do you really not understand who is running this scam:

".. populist political leaders are united by two claims: “(1) that a country’s ‘true people’ are locked into a moral conflict with ‘outsiders’ and (2) that nothing should constrain the will of the ‘true people’."

The GOP has been pretty direct in their messaging that they are about the "real Americans," and those are the MAGAs. Even to the point of this "gaffe"

Are you going to continue to defend the position we're talking about, because I'm not going to be distracted into a pointless tangent about whether "populist" should have quote marks here, or not?

We started with this quote, from the cited paper:

"Under this account, populist politicians can exacerbate cultural divisions, take a skeptical position towards science and expertise, and can be erratic in decision-making, eschewing the policy moderation that can result from between-party compromise."

In response you have wanted to play tin ear to what populism *and* decline of science mean.

"tin ear" I guess means "disagreed with me" to you.

And to note again, this is the literature review, not their hypothesis and something they claim is unsupported.

I've challenged you that survey doesn't show this and when challenged on this point you've hauled out some survey data which you claim showed changes within the right, but which don't, not predating "populism" but with what small changes there are following on from your boy Bush and his evangelical electoral alliance (e.g. if there is a Republican decline in faith in science, the pin is firmly on the pre-Trump Republican religious right, not "populism").

And while we're at it, we can dissect this paragraph:

Brossard says increasing concerns among scientists over science becoming partisan are not reflected in recent GSS polls, which show fairly modest differences between Democrats' and Republicans' confidence in the scientific community. While Democrats reported higher confidence in scientists than Republicans did in 2018, members of both parties have reported similar, high levels of confidence over the past 45 years.

First of all, a 45 year fish-eye lens is not going to capture the detail of a 4 year populist shift.

Second, if you do look back at past peaks and valleys, it's clear that something was going on. *Why* Democrats lost confidence in science in the 90's might be an interesting thing to discuss.

Rather than to dispense with as a "random walk."

Sure it might be interesting to think about why there were shifts in Dems in the 90s - at the extreme of their distribution more "hippies and hoteps" (and Unabomber types), then that filtering back into the centre?

But it's worth quantifying how much shift has happened and who has shifted, and how this sits in the historical context.

Like all the recent shifts, what's probably driving is a "Woke shift" of more radical and extreme shifts among a hardcore on the left who are heavily immersed in the internet, underway well before 2016 and at most mildly accelerated thereafter.

When multiple people start telling me I'm wrong, I start to wonder if I'm actually the person failing to follow the logic. Maybe everybody else isn't wrong, maybe it really is me?

"Under this account, populist politicians can exacerbate cultural divisions, take a skeptical position towards science and expertise, and can be erratic in decision-making, eschewing the policy moderation that can result from between-party compromise. "

And this is not a monopoly of the right. Indeed, "populism" is a word used by the left to describe policies of the right which the left would like to have.

Forget the fancy words: Nothing more than political competition in a democracy. :-)

This is not always true - 'Nothing more than political competition in a democracy.'

But forget about creating a simple definition that can be objectively applied. To give an interesting example - was Napoleon a populist? (Assuming one believes the French Republic to be democratic at any point, of course.)

Who wrote "The Death of Expertise: The Campaign against Established Knowledge and Why it Matters?"

And why did that author end up as a conservative anti-populist?

It strikes me that this book was written at a critical juncture, where there was still a possibility of a different, more science and expertise friendly, conservatism.

In page 39, the authors say

"To date, demographic and health-related variables account for 57% of the cross-national variation in per-capita Covid-19 mortalities."

I think those variables may account for at least two-thirds of the variation among relevant political jurisdictions. Any serious analysis focused on "the remaining" variables must start with a proper definition of the relevant political jurisdiction, a point of particular interest to large, decentralized countries like the U.S. Indeed, it's hard to define those jurisdictions.

The research referred to by Tyler is grotesque. At best, it's a good example of both the poverty of political science and academics' rampant opportunism. It's a pity that Tyler didn't warn us about how poorly the political and social variables were defined and measured for the empirical analysis. Please if you are going to read the paper pay attention to definitions and measurements. You don't need to know political science to question and laugh at the choices made by the authors.

What are COVID mortality figures worth? The attribution of a cause of death for a very ill, old, fat man is pretty arbitrary.

Or do they use Excess Deaths? Those are likely to be right but who's to know whether they are caused by COVID or lockdown disruption?

I haven't looked at the paper: how did they allow for the hypothesised differences in deadliness of the different strains of the virus?

I think I saw a report that half the Wuhan flu deaths reported in Minnesota actually listed something else as the cause of death on the death certificate.

It is mildly clear that in the US and Brazil, people are playing politics with the cause of death.

That's just as much reason to argue that covid-19 deaths are being undercounted, especially given that some cases are mild or asymptomatic and there continues to be under-testing.

I have no idea whether there's over-counting or under-counting. Not least because neither expression makes sense unless you conjecture that there is one single, true, knowable cause of death for every poor old geezer who dies. I think that pretty unlikely.

Having looked at estimates for seroprevalence and excess deaths to date for different English regions at points in time, I do tend to think very low (0.1%) and very high (3-5%) IFR on general "Western" demographics can fairly safely be excluded. Though of course there are questions about how many people are actually susceptible, but that's not exactly IFR.

I do tend to think the "lockdowns", in sense of the compulsory stay-at-home orders and compulsory business shutdowns, will prove to be a side issue, which will cause/prevent some deaths/economic damage, but will not be decisively factor. They weren't "The Hammer" and at best were a "soft tap". The main factor will be carehome protections, and those states who acted early / late on travel restrictions and social distancing, and the quality of understanding of where countries were in the epidemic (which relates to testing and skepticism of the WHO). And then simple luck at distance on travel routes to Italy.

"Lockdown" vs "Non-lockdown" will probably prove to be a distraction generated by the pointless rhetoric of the political side that noisily talks up their disregard for freedom and business and their regard for "public services" noisily exaggerating the benefits of lockdown (in protecting "public services").

Researchers must verify data reliability, but they don't' do it because it's too expensive. Governments and bureaucrats usually manipulate the data because they know others cannot verify them (when was the last time you reviewed the manual and the protocols to enforce the rules?).

The public discussion of data about the virus, the disease, and the pandemic has shown how unreliable they are in all countries (I'm not surprised the authors excluded China but I'm not sure that they did it because of its unreliable data). My reference to academics' rampant opportunism in my previous comment is partly related to the use of unreliable data.

"academics' rampant opportunism": very true. Thank God scientists are not subject to that, nor economists.

It reads like the usual social science thinking that groups (states) capacity control everything. In reality, the Ro term in the epidemic models is a pure socially determined variable that is the product of the frequency of human interactions and the probability of transferring the virus for each type of interaction.

It is individual behavior that determines the probability of transfer on each interaction and groups of people of all categories are overlapping distributions of behaviors determine the transfer probability. A health care worker with PPE can work up close with an infected individual and have a reasonably low probability of transferring the virus but Joe citizens without a mask may see a transfer with even some distance between the infected person, who may or may not have symptoms, and the individual.

As a practical observation, once the number of infected becomes larger than what can be tested, traced, and isolated the virus becomes free-running and the behavior becomes a function of the local Ro of individuals and small groups. Shutdowns of large groups, transportation systems, etc. have little impact at that point and only individual behavior like wearing masks (sanitized between use), washing outer surfaces (hands, face, hair, and outer garments that may have been contaminated will reduce the individual Ro and protect the individual.

State competence or incompetence can impact that point of where the virus becomes epidemic and uncontrollable. Our state monopolies like the FDA prevent testing to keep their regulator fiefdoms in the loop for at least 30 days at a time with 3-day doublings allowing the problem to become 1000 times larger and beyond the ability to control via testing, tracing, and isolating.

I am working on some similar research on variation in COVID-19 death rates between countries, looking at various factors other than government intervention, such as population characteristics and culture. Have not written up a paper yet but did summarize what I have found so far in this article. https://www.japantimes.co.jp/opinion/2020/06/12/commentary/japan-commentary/japans-low-covid-19-death-rate-due-higher-cultural-level/#.Xuh8kOeRVPZ

Comments for this post are closed