Welcome to Plugging the Gap (my email newsletter about Covid-19 and its economics). In case you don’t know me, I’m an economist and professor at the University of Toronto. I have written lots of books including, most recently, on Covid-19. You can follow me on Twitter (@joshgans) or subscribe to this email newsletter here. (I am also part of the CDL Rapid Screening Consortium. The views expressed here are my own and should not be taken as representing organisations I work for.)
Today’s post is part book review, part an introduction to economic theorising and part a critique of public health decision-making. Let’s do the book part first. I have just finished reading Michael Lewis’ The Premonition: A Pandemic Story. I found it completely eye-opening. Basically, it was all pretty much new to me and I consider myself someone who has pretty well-read with regard to Covid-19 and pandemics. Lewis has a knack for finding the people who are really interesting during crises. He has mostly done this for financial crises but this time we had a public health one. True to form, these are the people who saw what was happening before anyone else, did not have the power to do something but pushed forward regardless. Read this book. You won’t be disappointed.
Let me give you an example of something that I did not know prior to reading this book: the entire concept of social distancing as a way to manage pandemics has only been accepted by epidemiologists and public health people for 15 years. You read that right, a decade and a half. Prior to 2006, the main playbook for a pandemic was to just ride it out and try and get treatments and vaccines quickly. I confirmed all of this with an epidemiologist colleague. I was aghast. Lewis runs through the story of how their minds were changed. It was a combination of theoretical computer modelling from outside the field along with a re-examination of 1918 that showed that social distancing helped keep pandemic fatalities low. (The same research that I used to discuss the issue in my own book).
First-order effects
You might be wondering “how could this be?” Isn’t it obvious that social distancing is the way to go. The answer is ‘yes’ it is obvious. And it is something that is baked into how an economist theorises that may well not be as common as we might have thought.
I’m primarily an economic theorist which means that I think about making predictions that are based on clear assumptions and logic. This is both useful as a way of framing data and of making decisions when there is no data. But there is an art to it because you have to abstract from reality and make sure you capture what is important.
The first task, therefore, is to make sure your model accounts for first-order effects. A first-order effect is a key insight from the model that arises in all cases. A good example of this in economics is that people cannot consume more than has been produced of a good. It is pretty obvious but if you are creating a model you need to make sure you have explicitly accounted for it. It doesn’t just happen but when you have it there it is always present.
There is an equivalent insight for a pandemic. A pandemic is literally something that impacts on all people through person to person transmission. That means that the following is fundamentally true:
IF YOU ISOLATE EVERY PERSON FROM EVERY OTHER PERSON FOR A SUFFICIENT TIME, THE PANDEMIC WILL END
You can’t have a pandemic on your own. Other people are required.
If you understand this, you can understand why it was so shocking that social distancing wasn’t considered a tool of pandemic management prior to 2006. Social distancing takes this fundamental truth literally and aims to isolate people from one another. How could this not be part of the tool kit? After all, we know that people, when they understood others were risky to them, would stay away from each other. But why wasn’t this a policy option?
Lewis covers that turf. But the bottom line is that experts did not believe social distancing could be mandated. They either said people wouldn’t do it or they believed it would be too costly for people to do and so took it off the table for that reason.
Now those issues could be true and, let’s face it, throughout Covid-19 both issues have come about in waves. But each of these issues are not first-order but second-order effects. They apply in certain cases. What was extraordinary about the choice not to use social distancing as a policy tool was that those choosing had no information about whether those cases were reasonable or not let alone whether they would apply uniformly.
The point about a first-order effect is not that second-order ones don’t exist but that it is highly likely you will obtain insights and impact from the first-order effect somewhere. And you are more likely to have that impact if you have an institutional framework that at least admits the possibility that it is a tool you might want to use.
The Information Problem
Another thing that Lewis’ account of social distancing gave me insight into is something that has puzzled me through Covid-19. How was it that, I, a decidedly non-expert in pandemics in March 2020, was able to come up with a way of framing the policy issues that were not being talked about by most epidemiologists and, indeed, resisted by them? The answer was that epidemiologists were repeating their earlier mistake.
My insight was the following:
IF YOU IDENTIFY AND THEN ISOLATE EVERY INFECTIOUS PERSON FOR A SUFFICIENT TIME, THE PANDEMIC WILL END
This insight takes the first-order insight above and qualifies it. From an economics perspective, it is pretty close to first-order but it may fail the ‘all cases’ requirement because there may be viruses where it is impossible to identify infectious people. I don’t want to quibble about that. The point is that it was an insight that informed a policy direction. The critical feature of it was that it was a policy direction — investing in the information we needed — designed to improve on the fundamental pandemic insight that we need to isolate everyone. (There was some understanding of this in the literature but really in the context of traditional measures rather than a re-framing of the problem).
But again this was a set of tools that were not part of the playbook in 2020 — unlike social distancing which, thankfully, was by then. Why wasn’t it? There were arguments such as unavailability of tests, the costliness of contact tracing, privacy rules and that people would either not isolate or behave badly if they knew if they were infectious or not. The kind explanation is that we were socially distancing and didn’t need to do better. The reality proved differently of course. Just because isolating everyone could resolve the pandemic, apart from Wuhan, no government actually did that.
My point is that all of the objections were second-order. They might apply but then again they might not. And we did not have the information relevant to adjudicate that. This is not a good enough reason to take a policy tool from the table. You would have thought that having already learned that lesson in recent memory, the epidemiology and public health fields would be more open to other tools. But the sad experience is that this has not happened … yet.
From a reasoning perspective, the difference between first-order and second-order effects is that the former can rest on logic while the latter requires data. By ignoring first-order effects, they were throwing away logic but did so without verifying what replaced it with data. Suffice it to say, that is a bad way of formulating policy responses and evaluating interventions.
Resistance to social distancing
But I want to come back to social distancing and why it was resisted as an option pre-2006. Remember this was after the SARS scare. Interestingly, the conventional wisdom regarding that scare and why it did not go global was that people were symptomatic before they were infectious and so isolation could happen quickly. To my mind, that was a reflection of the information problem. But what about isolating everyone?
In 2009, this seemed to be the consensus opinion:
Measures aiming to increase social distance, like cancelling mass gatherings or closing businesses, had only little additional impact on reducing Re of SARS when applied in addition to the core control measures of early case detection, hospital infection control, tracing of HCW and of close contacts (Fig. 2). Apart from the low efficacy, these measures strongly interfere with daily life as was seen in Beijing/China, where traffic reduced by 75% leading to adverse social and economic consequences (Balasegaram and Schnur, 2006).
The last sentence interested me. The statistic was measured at one station in Beijing (Western Station; the main rail terminal). But this wasn’t a mandatory social distancing. Instead, the public health measure was to increase surveillance by people of each other for symptoms. People took it upon themselves to distance from one another. But the “adverse social and economic consequences”? There were some estimates that the whole event cost $60 billion but the sources of that seemed to be an impact on tourism which is very different from adverse social consequences.
With respect to SARS, precisely because it was short-lived, no one thought to add to the toolbox of interventions to include social distancing. What’s more, it wasn’t actually done during SARS. In the end, it doesn’t appear to have been seriously considered even though there is evidence that ordinary people acted on it anyway. That alone would have provided data to inform on second-order effects but there appears to have been nothing. (I am happy to be corrected about any of this as I tried to find things but may well have missed something.)
The Bottom Line
In the end, Lewis’ book provides a broad swath of some of the answers to what was going on. There are several heroes in the story but what is really important is to understand why the heroes weren’t in charge when they needed to be. This seems to have been the pandemic experience in so many places. What I would love to hear about is the possibly more boring story of how the heroes that existed in places that did manage the pandemic properly ended up where they needed to be.