New York magazine has a good article on Covid-19:
“Basically, going back to January, they’d be like, ‘China’s not going to control it; 80 percent of the population is going to get it; all efforts to contain it are going to fail; we have to learn to live with this virus; contact tracing and testing make no sense; this is going to be everywhere; right now we need to build up hospitals’ — which they didn’t even do. But they really didn’t think it was stoppable,” she says. “And then all of a sudden you started to see, in February, South Korea stopping it, Taiwan stopping it, and China stopping it. Then, in March, New Zealand. And then Australia. And then there’s this realization of, ‘Oh, wow. Actually, it is controllable.’”
At the beginning of March, South Korea was averaging more than 550 new daily confirmed cases, compared with just 53 in the U.K. At the end of the month, South Korea had 125; the U.K. was at 4,500 and climbing. “In the UK we have had nine weeks to listen, learn and prepare,” Sridhar wrote angrily in the Guardian, berating the British regime for failing to establish basic systems for supplies, testing, and contact tracing.
Later they point out that things are not quite that simple:
Francois Balloux, an infectious-disease epidemiologist and computational geneticist at the University College of London, goes further. “It’s not obvious that different measures taken in different places have clearly led to different outcomes,” he says. “There’s a lot of idiosyncrasy, and I think it’s simplistic to say that the countries that have controlled or eliminated the virus did things extremely differently. If you just list, for instance, the interventions that places like New Zealand or Australia have implemented, they’re not drastically different — in stringency nor duration — than in some other places. The country that had the strictest lockdown for longest in the world is Peru, and they were absolutely devastated. I think the slightly depressing message,” Balloux says with a sigh, “is that there is not just a set of policies that will bring success and can just be applied to any place in the world.”
So how can we reconcile these two conflicting narratives? First we need to distinguish between public policy and behavior. I suspect that the relatively low level of Covid deaths in some areas of the US (Washington, Oregon, Utah, Northern New England and even the SF Bay area of California) has more to do with culture than public policy. People behave differently in different parts of the US. If death rates in the Pacific Northwest and northern New England are similar to those in Canada, is it so far-fetched to believe that their culture also resembles Canada more than it does much of the rest of the US?
But the big international differences may require an additional explanation. Reading the article, I was immediately reminded of the global recession of 2008-09. I’ve argued that the recession was caused by tight money policies, especially in the US and Europe. But why was Australia able to avoid a recession? Their central bank didn’t do any QE, and didn’t even cut interest rates to zero.
In fact, what to the average person looks like an “easy money” policy is often the exactly opposite. It’s precisely because Australia had a more expansionary policy early in the recession that they were able to avoid some of the more “reactive” policy measures employed elsewhere during the 2010s. Similarly, the US was a bit more (proactively) aggressive than the ECB during 2009-10, and as a result the ECB ended up being forced to do aggressive (reactive) QE and negative interest rates in the middle 2010s.
So if you see news stories of positive interest rates in Australia during the global recession of 2008-09, do not conclude that easy money is not stimulative. And if you see news stories of restaurants being open in Taiwan, Australia and New Zealand during the Covid pandemic, do not conclude that social distancing is not helpful. Rather the positive interest rates are a sign that Australia took proactive steps to prevent a deep fall in NGDP growth, and the open restaurants are a sign that they got on top of the pandemic early on, with an aggressive policy aimed at driving Covid rates down close to zero.
There’s another interesting comparison between Covid and the 2008-09 recession. In both cases, bloggers were often ahead of the experts in diagnosing the problem and recommending appropriate policies. Bloggers pointed out that the Fed’s October 2008 decision to begin paying interest on reserves would have a contractionary effect. Today, that criticism is widely understood as being correct. Indeed in his memoir, Ben Bernanke acknowledges that monetary policy was too tight after Lehman failed. Similarly, bloggers like Alex Tabarrok and Tyler Cowen have been consistently right in their criticism of the public policy response to Covid.
PS. The US is currently at 1670/million Covid deaths. Canada is at 595/million, or halfway between Utah and Oregon. Here are the lowest 7 states:
Note: The 15 highest Covid death rates are in both northern and southern states, as well as both urban and rural.
READER COMMENTS
Brett
Mar 23 2021 at 10:52am
Utah doesn’t surprise me. At least in Salt Lake County, the state government was slow to do a mask mandate, but people were much more quick to almost universally wear masks indoors.
John Alcorn
Mar 23 2021 at 11:30am
USA vs EU/EEA/UK is the aptest comparison for pandemic outcomes. Both are vast, modern, wealthy, western federations (of States or Nations). Each federation exhibits mainly decentralized pandemic control policies, and a mix of centralized and decentralized pandemic relief (and vaccination) policies.
A striking fact: cumulative Covid-19 death rates in the USA and in the EU/EEA/UK are roughly the same. The figures are 1.6 per thousand (USA) and 1.4 per thousand (EU/EEA/UK). (The comparison changes little if you omit the UK.)
Dynamics of behavioral adaptation (including also pandemic fatigue) and correlative dynamics of government action — What Tyler Cowen calls the “epidemic yoyo” — help to explain why cumulative pandemic death rates have been roughly the same in these two huge western federations.
A question arises: What, then, explains more substantial variance, among European nations or among American States, within each federation? Scott Sumner invokes culture, which indeed must have bite.
John Ioannidis (Stanford U.) has pinpointed a different crucial cause of variance, at least in the pandemic’s first wave; namely, differences in eldercare institutions and practices. See his recent study, “Precision shielding for COVID-19: metrics of assessment and feasibility of deployment,” BMJ Global Health (6 January 2021): https://gh.bmj.com/content/bmjgh/6/1/e004614.full.pdf
Ioannidis provides a scorecard. What we still need, to explain the pattern in the scorecard, is a full upstream account of institutional differences among nations and among States. For example, if I understand correctly, Norway’s eldercare institutions are small and compartmentalized; whereas Sweden’s are large and are staffed by marginalized immigrants — about a quarter of Sweden’s eldercare staff were infected during the first wave of the pandemic!
Here is the abstract of Ioannidis’ fascinating article:
raja_r
Mar 23 2021 at 11:51am
India’s death rate per million is around 110/million.
While the Indian central and state governments’ policy might have been aggressive, the compliance from the people was abysmal (if you’ve seen the traffic in Indian roads you know what I mean). In my hometown, the compliance was so bad that the police were beating people with sticks to get them to stay in their homes. They gave up after a few weeks.
Czechia’s death rate is around 2300/million.
My wife is from the Czech Republic. They implemented a strict lock-down – my father-in-law could not visit his 92-year old mother for several weeks in March/April 2020. They had much better testing, mask mandates, and compared to India, the compliance with the government policy was very high (no police beat-downs were needed).
In India’s case, the low death rate wasn’t because of social distancing, contact tracing, a culture of compliance or an excellent health system.
In Czechia’s case, the high death rate was not due to a lack of social distancing.
There are more variables involved (urban vs. rural, demographics, the initial seeding from travelers, etc.). I think a narrative of “good vs. bad response”, or a “proactive vs. reactive response” is too simplistic.
(Even cities like Chennai & Mumbai, which have a similar population as Czechia, have a lower death rate.)
Scott Sumner
Mar 23 2021 at 11:54am
Some people have argued that Indians may have some natural immunity due to previous viruses. I have no firm opinion on that question.
robc
Mar 23 2021 at 1:20pm
At least one study makes it clear the virus primarily targets old, rich, and fat “countries”. There were lots of cross-variables, but basically countries where people don’t die young from other viruses were more likely to have high covid deaths. I don’t think necessarily due to immunity due to other viruses (although that seems to be part of it with the very young vs old) as much as more unhealthy targets. If few live to be 60, then there are less people over 60 to die.
The “rich” is obviously not a real effect, just one that correlates well with the others.
zeke5123
Mar 23 2021 at 4:22pm
If you allow for the possibility of cross-immunity, then isn’t the cross sectional differences between pretty much all jurisdictions potentially explainable by pre-existing immunities? Of course that doesn’t “prove” anything. But it is hard to say that Cowen or Tabarrok was correct if we don’t really understand why some policies failed and some succeeded.
Scott Sumner
Mar 23 2021 at 5:45pm
We certainly know enough to know they were right about vaccines, and I suspect they were right about testing and a wide range of other issues.
E. Harding
Mar 23 2021 at 6:33pm
Indian deaths are obviously understated by a factor of more than ten. Lower-middle-income countries cannot be relied on for accurate COVID statistics.
raja_r
Mar 23 2021 at 10:45pm
“Some people have argued that Indians may have some natural immunity due to previous viruses.”
Exactly. Like I said, there are more variables involved here than just social distancing, lock-downs, government policies and culture.
Travis Allison
Mar 23 2021 at 12:28pm
@raja_r Do you think the statistics about the death rate from covid in India can be believed? Could the numbers be off by 10-20x? I have no idea how good the health system is at recording cause of death and/or whether there was political pressure to not record covid as cause of death.
raja_r
Mar 23 2021 at 10:42pm
“Do you think the statistics about the death rate from covid in India can be believed?”
As much as you can believe numbers from China and other countries. There are too many state governments, municipalities and other health authorities for them to all act in unison and cover-up.
I believe the numbers from European countries and the United States are wrong in the other direction – many of the deaths are “with Covid” rather than “from Covid”. I read about instances where suicides were counted as Covid deaths. That might explain some of the difference.
“Could the numbers be off by 10-20x? ”
No. Even 3x is a stretch.
Max More
Mar 24 2021 at 12:24pm
But 220% is not at all implausible. If what I’ve read is correct, well over half the population of India is rural and deaths there are not reliably recorded and count just about not at all in the national COVID death counts.
K.D. Walter
Mar 23 2021 at 12:25pm
The conclusion I drew from the success of Taiwan et al was that rigorous travel restrictions put in place early were the most effective policy by far.
Whether those successes could be replicated in a non-island country is of course an open question.
Zach
Mar 23 2021 at 12:58pm
Utah and Alaska have a demographic advantage: they have the two lowest median age statistics among the 50 states, respectively. However, Maine and Vermont have some of the highest median age statistics, which lends credence to your culture theory. I suspect there is a bit of both at play.
I am from rural Utah but now live in the Wasatch Front. I can attest that there was greater adherence to social distancing measures in the latter than in the former (though it has consistently been getting more relaxed in both areas). But the high number of young people has shielded us from a high number of severe cases.
Scott Sumner
Mar 23 2021 at 5:47pm
I am visiting Arizona, and I can see why they have more deaths/person than California. The people here don’t seem as cautious.
Max More
Mar 24 2021 at 12:30pm
Scott, I can’t speak to your personal experience in making that comparison. I’ve lived in Arizona for ten years (although in Scottsdale, which is not at all representative) and have observed a lot of resistance to restrictive policies, at least early on. In recent months, mask-wearing (which I’m increasingly skeptical about) has been around 77%, or similar to most states.
Arizona does have one of highest percentages of Hispanics and that is a known risk factor. It also has a relatively high population density (most of us live in Maricopa County). To a lesser degree, we have a relatively high percentage of old people and people in nursing homes.
So there are many other factors at work. Also, I’ve observed rapid increases and decreases in cases which have no correlation whatsoever to people’s behavior. Just look at our enormous plunge in cases since January 11. We went from one of the fastest-growing states to slowest-growing (and lowest estimated infection rate), yet there was no change in behavior.
Scott Sumner
Mar 24 2021 at 2:31pm
Even with no change in behavior, R can fall due to greater “herd immunity”.
Max More
Mar 27 2021 at 8:23pm
True. But I don’t see how that relates to what I said. I was pointing to factors other than cultural differences that could explain relatively higher or lower confirmed case rates and death rates.
Alan Goldhammer
Mar 23 2021 at 1:00pm
Unquestionably personal responsibility had a lot to do with control of COVID-19 spread. Early adopters of mask wearing coupled with appropriate physical distancing tended to make it through in good shape. No secret about any of this. Most of my friends who were colleagues of mine from the pharma industry knew what the severity of this would look like back in mid-February and we all adjusted our behaviors. I don’t think there was a single person among our email list that came down with COVID-19.
John Alcorn
Mar 23 2021 at 1:45pm
Numerous studies find that essential workers incurred much greater risk of severe COVID-19. See, for example, Evangelia Demou et al., Occupation and risk of severe COVID-19: prospective cohort study of 120 075 UK Biobank participants,” BMJ: Occupational and Environmental Medicine (9 December 2020). Let’s keep personal responsibility in perspective.
Some individuals are in fact in a position to make decisions that have a major impact of prevalence of acute cases; for examples, Governors who decide whether to place COVID-19 patients in eldercare facilities, instead of providing focussed protection for the most vulnerable demographic. Personal responsibility writ large!
Kailer
Mar 23 2021 at 1:54pm
I think the second guy has it right. How else do you explain the whole Asia-pacific doing well, despite quite a bit of heterogeneity in policies. Noticeably absent is from this discussion is Japan. Japan was in denial about COVID and hardly did anything. I agree that Tyler Cowen has been right about a lot of the pandemic, but he himself would admit to getting Japan wrong, and they’re monetary policy sucks.
There’s a lot of post-hoc reasoning in these types of comparisons. You start with the outcome and then backwards justify it based on what policies they did.
Scott Sumner
Mar 23 2021 at 5:50pm
I do not agree about Japan. Their culture is much more likely to follow norms on mask wearing, avoiding hugs, etc.
Compare Hubei province with the rest of China and it’s obvious that behavior plays a huge role.
Todd Kreider
Mar 24 2021 at 4:45pm
The last sentence is true.
Japan declared an emergency for large cities (40% of the population) on April 8 and then extended that to the entire country April 16 but cases were already falling before they could have had much or any effect.
Japan also was at the bottom of the OECD in testing with South Korea second to the bottom. Unlike South Korea, most of the testing was not track and trace so the media almost never includes what happened in Japan.
Todd Kreider
Mar 24 2021 at 4:47pm
(That is, “You start with the outcome and then backwards justify it based on what policies they did” is true. and unfortunately rampant.)
Max More
Mar 27 2021 at 8:31pm
Interesting. It remains true that Japan and South Korea have much lower levels of testing than the USA. Using current numbers from Worldometer (isn’t the modern world of data availability — if you can trust it — wonderful?):
Current total test rates relative to USA:
Japan: 6.3%
South Korea: 12.3%
sean
Mar 23 2021 at 2:13pm
Regardless, once you realize you society will fail to complete eliminate the virus your decision tree changes signficantly.
Island Nation (or ability to keep virus out once gone) – hard lock down for 6 weeks then party
America – 1 + year of restriction or ignore it. choose ignore it and take the deaths.
Non-homogenous populations without strong borders aren’t capable of limiting spread. If you know its going to fail might as well just live instead of having the worst of both worlds.
Jon Murphy
Mar 23 2021 at 2:23pm
The choices aren’t quite as stark as you present. Australia and New Zealand have essentially been whipping back and forth between lockdown and reopen. Despite being island nations, they cannot keep the virus out
robc
Mar 26 2021 at 8:41am
Also, New Zealand apparently is having hospital overload issues. Not due to covid cases but due to covid policy. You know, the exact thing that the policies were originally sold to prevent.
Michael Rulle
Mar 24 2021 at 11:17am
One idea you do not mention is that differences may just be random. There are many components that could be random. We do not know enough to do more than speculate as you do in this essay. Plus our data is really not very good.
I like speculation, but it should be recognized as such. For example, the 2 month period a year ago in NY/NJ cannot me explained with evidence. But 6-7x excess deaths is astounding—-but not easily explained—-I think we cannot explain it.
If you think the northern borders of the US are culturally similar to Canada, you have not spent much time in Toronto or Quebec, or the Canadian oil belt.
I worked for almost a decade for a large Canadian institution. I went in thinking our cultures were the “same”. Big picture, of course, they are similar. But to suggest at the micro level of “Covid behavior” it represents a similarity of culture is really just a wild assertion.
Scott Sumner
Mar 24 2021 at 2:36pm
The NY/NJ experience last March is easy to explain. The virus came into the region in people returning from Italy, and there were no precautions being taken at the time, so R0 was very high. No different from Wuhan during the early stages of the pandemic.
I disagree about Canadian culture. My experience is that it’s more similar to states like Washington than to states like Mississippi. I never said it was identical.
BW
Mar 24 2021 at 2:45pm
That’s huge, if true. I haven’t been swayed by the anti-lockdown arguments, thus far, because of the “but China” counter-argument. The “but China” counter-argument: the Chinese should have millions of covid deaths, given their population; but they only have thousands. Either they managed to hide millions of dead bodies (not likely), or their lockdown worked. Given that the Chinese enforced their lockdown far more strictly than every other country, I just figured that lockdowns didn’t work elsewhere because they weren’t strong enough.
But if Peru had a lockdown more strict than China, and it didn’t work, then that would change my mind. Does anyone have any sources providing details about the strictness of Peru’s lockdown?
Todd Kreider
Mar 24 2021 at 6:03pm
“the Chinese should have millions of covid deaths, given their population; but they only have thousands. Either they managed to hide millions of dead bodies (not likely), or their lockdown worked.”
China with one lockdown and unreliable data has had 69 Covid deaths per million. Japan, where the government encouraged social distancing after cases had start to come down (same for the lockdown countries like the UK, Spain and Italy) has had 71 Covid deaths per million. People in both Asian countries have very low obesity and almost certainly had much higher prior immunity than the West.
Czechia 2,350 deaths per million; Hungary 1,970; Belgium 1,960; UK 1,850; Bosnia-H; 1830; Italy 1,760; Slovakia 1,700; U.S. 1,680; Portugal 1,650; Spain 1,590; Mexico 1,530; Peru 1,520 deaths per million.
So Peru is the 12th highest in Covid-19 deaths per capita with strict lockdowns and Sweden is around #15 with no lockdowns and little mask use – not that masks mattered anywhere.
Todd Kreider
Mar 24 2021 at 8:14pm
Oops, I read the wrong column. China, whose numbers I don’t trust, claim to have had 3 Covid deaths per million. It might be that low but may not be.
Max More
Mar 27 2021 at 8:39pm
Todd: Of course any sensible person would not take China’s reported numbers at face value. Much more worrying: Given the nature of the regime, even if the numbers were remotely accurate, how do we know this result was not achieved by rounding up potentially-infected persons and either incarcerating or murdering them? Even today in the US, it’s hard for us to take such a scenario seriously but anyone who has studied the history of totalitarian and authoritarian regimes should at least take such a scenario seriously.
AMT
Mar 27 2021 at 12:21am
Something I thought is obviously critical just from the title of the post, but not discussed so far is the timing of the policies implemented. Maybe huge differences in outcomes are mostly dependent on how quickly policies were enacted, rather than the severity.
E.g. Are we uncertain how helpful water is at putting out house fires, because we lament that one house suffered far more severe fire damage despite a far larger amount of additional water poured on it, compared to another house that was reached far more quickly?
Comments are closed.