Sunday , January 17 2021
Home / Scott Sumner /Non-linearities in Covid outcomes

# Non-linearities in Covid outcomes

Summary:
Recent trends in Covid-19 fatalities in Western countries are quite unusual, with a wide range of outcomes. We know that these highly divergent results can be explained with a model where long run outcomes are highly sensitive to whether the replication rate “R0” is above or below 1.0 (after social distancing.) I will argue that a country’s complexity plays an important role in determining that replication rate. Obviously the term ‘complexity’ will require some unpacking, but first let’s look at the number of Covid deaths thus far in November: EU: 34,276 deaths (76.56 per million) USA: 14,637 deaths (44.12 per million) Canada: 658 deaths (17.38 per million) Australia: Zero deaths (0 per million) New Zealand: Zero deaths (0 per million) I will argue that in the list

Topics:
Scott Sumner considers the following as important: , , , , ,

This could be interesting, too:

Scott Sumner writes The “mad scientist” problem

David Henderson writes Noubar Afeyan on Academia, Business, Immigration, and the American Dream

Scott Sumner writes There’s no one behind the curtain

Scott Sumner writes Biden approves First Dose First!

Recent trends in Covid-19 fatalities in Western countries are quite unusual, with a wide range of outcomes. We know that these highly divergent results can be explained with a model where long run outcomes are highly sensitive to whether the replication rate “R0” is above or below 1.0 (after social distancing.) I will argue that a country’s complexity plays an important role in determining that replication rate. Obviously the term ‘complexity’ will require some unpacking, but first let’s look at the number of Covid deaths thus far in November:

EU: 34,276 deaths (76.56 per million)

USA: 14,637 deaths (44.12 per million)

Canada: 658 deaths (17.38 per million)

Australia: Zero deaths (0 per million)

New Zealand: Zero deaths (0 per million)

I will argue that in the list above, countries with higher recent death rates are places with higher levels of complexity. And I’ll also argue that a slight difference in complexity can make a huge different in long run outcomes. And finally, I’ll argue that these results can be affected to some degree by policy choices, but mostly for countries near the “tipping point” (i.e. places like Canada and Australia.)

Before going further, let me address the concern that these results only show recent rends, and thus for instance the US has been hit harder than Europe if you look at the entire pandemic, not just November. Or that Australia and New Zealand had some deaths before November. That’s all true, but I’m interested in current trends because I feel they better illustrate the direction to which countries tend to migrate in the long run.

There are many possible reasons why Australia and New Zealand might have done better than other Western nations. For instance, Australia does not have particularly cold weather. But you could say the same about Texas, which had over 200 deaths yesterday. Or perhaps Australia was just lucky; the virus missed this remote continent.

But the Melbourne area was hit by a huge surge in cases a few months ago, with hundreds of new cases every single day during July and August. Perhaps they avoided “superspreaders”, but how likely does that seem when total cases are in the tens of thousands? There’s the “law of large numbers” to consider. How was Australia able to get things under complete control in a short period of time, and why weren’t other Western nations able to replicate that success?

Consider a model where Covid is easiest to control in an isolated village of 100 people, where everyone knows each other. As societies become more “complex”, Covid becomes progressively more difficult to control. But what exactly does the term  ‘complexity’ mean in this context?

I’m open to suggestions, but I’d start with density. Next I’d add the total population of a country. Then I’d add the ease of movement between population centers. Highly populated and dense countries with lots of movement between regions are highly complex.

Then I’d add cultural heterogeneity. That factor may be negatively correlated with civic cohesion, or willingness to cooperate for the public good. You might want to add administrative complexity; are the governmental lines of authority clearly demarcated?

Here’s another way to make the distinction. Travel in New Zealand is both much more convenient and much less interesting than travel in Italy. Italy is complex, while New Zealand is “simple” (no pejorative intended.) I’ve lived in both the UK and Australia, and Britain seemed like a much more complicated and confusing country. Less “legible” if that term has any meaning when applied to countries. I suspect that the UK’s greater density plays a big part in that difference. And notice that while hard hit Belgium is a small country, it’s also quite densely populated and culturally diverse, with a confusing governmental structure.

Although Australia has a population roughly comparable to Texas, and also has some metro areas that are only a bit smaller than Dallas and Houston, it differs in one important respect. The Australian population centers are more isolated than in Texas. In a sense, Australia is sort of like five New Zealands cobbled together—with population centers that are pretty isolated from one another by vast distances. People don’t typically just get in the car and drive from Adelaide to Perth. So when commenters tell me what Australia did differently, such as interstate travel bans, I want you to also reflect on the extent to which these policy differences are partly endogenous, reflecting geography and culture.

You might argue that Canada is kind of similar to Australia, both being continental size English-speaking countries with modest populations. But Canada is more diverse, with a French area that was hit far harder than the rest of Canada, including more than 60% of Canada’s Covid deaths. Right now, the four Maritime Provinces have a grand total of 43 active Covid cases, while Quebec has 13,463. Canada may also have more links to the US, despite recent travel bans.

In this model, even a slight difference in complexity can have big long run consequences if it puts two countries on the opposite side of R0 = 1.0. Canada had the misfortune of having a bit too much complexity to control Covid (or perhaps a bit less effective government policies). Over time, the two countries diverged more and more, with Australia going to zero deaths and Canada to a position somewhere between Australia and the much more complex US/EU regions.

The big policy question going forward is whether in a future global pandemic there is a set of policies that if pursued early and aggressively could get us to the Australian equilibrium. I don’t believe that any one policy could do that for the US or the EU, but I wouldn’t rule out a set of policies in combination. These would include a much earlier travel ban from the country where the virus originates. And a much more aggressive test-trace-isolate regime for the few cases that sneak though the travel ban.

It’s much easier to control an epidemic if you don’t first allow it to get out of control, but (and this is important) Melbourne showed that it’s possible to eliminate a pandemic even after it’s out of control. That’s very good news.

My suggestions might lead to an overreaction to less serious threats, such as the earlier SARS virus from 2003. But in a sense what I think doesn’t really matter. The reality is that future SARS-type outbreaks will be accompanied by some pretty draconian travel bans, at least until scientists can figure out the exact risk associated with the new virus. That’s the new world we live in, for better or worse. And for the few cases that do sneak through, expect countries to try very hard to replicate what Melbourne did.

PS.  I hope it goes without saying that I am not recommending that countries become less complex.  Complexity also confers huge advantages.  It helps explain why industries like Hollywood and Silicon Valley locate in the US rather than New Zealand.

PPS.  When examining the following graph, pay attention to the log scale:

Scott B. Sumner is Research Fellow at the Independent Institute, the Director of the Program on Monetary Policy at the Mercatus Center at George Mason University and an economist who teaches at Bentley University in Waltham, Massachusetts. His economics blog, The Money Illusion, popularized the idea of nominal GDP targeting, which says that the Fed should target nominal GDP—i.e., real GDP growth plus the rate of inflation—to better "induce the correct level of business investment". In May 2012, Chicago Fed President Charles L. Evans became the first sitting member of the Federal Open Market Committee (FOMC) to endorse the idea.