The Covid blame game
Here, I discuss a new book by Debora MacKenzie, COVID-19: The Pandemic that Never should have Happened and How to Stop the Next One that was just published. How should we evaluate the decisions that were made?
Who is to blame for our current predicament? Yes, I know, it is not really the time to play the blame game when we are still well in the midst of a crisis. After all, does it really matter right now how we got here? And if there is someone to blame, then what? How does that help us right now?
But we can’t help ourselves and keep those questions just rhetorical for the moment. For most of us, the crisis has left us time to think and reflect. And how we got here is a big question and as this week’s article from Ed Yong on the US in The Atlantic shows, we are very interested. Debora MacKenzie, in her new book, provides the information we need to consider that question and it is worth your time.
Debora MacKenzie has been a journalist writing for New Scientist on infectious diseases for decades. This is her first book although you would never know it given its breezy yet authoritative style and the fact that it was written in two months which is an unprecedented speed for anybody, present company excluded. It is, by far, the best book written about Covid-19, present company included. There are so many interesting issues that this will not be the last post I write based on the book.
MacKenzie takes a look at the question of whether all of this can be avoided. Her answer is a deep one: with present institutions, no it could not have been avoided. To avoid pandemics, we need institutional reform that is clear and, surely, doable.
Today, I am going to focus on the present institutions. One of the things that frustrated me when writing my book was that it was hard to get a handle on precisely what everyone was supposed to do. We had national governments, regional governments, quasi-independent authorities like the CDC and then international institutions like the WHO. We had scientists. We had warnings. We had people with experience. Sure, we had politicians that were struggling but we also knew we would have them. It wasn’t a surprise. Very little was a surprise to people paying attention to these matters and there were plenty of them.
One of the advantages of thinking about these questions now is that it is easier to remember what we didn’t know when key decisions were being made. In my chapter on “Predictable Surprises” which was written in the same month as the lockdown decisions were taken, I freshly recalled the week prior to those lockdowns when increasing numbers of people understood the mathematics associated with the pandemic and that very small case counts could balloon very quickly through the power of exponential growth (as infections beget infections) to an out of control medical emergency. If you take action a little sooner (say a week), you have an order of magnitude fewer cases and fatalities. The days waiting for leaders to take action seemed an eternity but around the world, those days were waited.
[Before going on, one brief interlude to point to new research from Richard Holden and DJ Thorton that takes the baseline SIR model and adds properly-specified uncertainty so that the reproduction rate of a virus is not a certain and predictable number but itself is uncertain. Because infections drive infections, this means that two regions that start from EXACTLY THE SAME PLACE with regard to the virus and do EXACTLY THE SAME THINGS can have very different experiences. In other words, whether you have a mild pandemic or a serious one. Put simply, the reproduction number (R_t) has a variance that is a function of the proportion of the population who is susceptible to the virus. In other words, the path of the virus is most uncertain when the pandemic is just starting. This means that while you can try and push R_t below 1 (so that the virus starts to die out), the fact that R_t can jump around means that you have to really push at it. Of course, this is just another reason to contain viruses early and often; as if we needed more reasons! But this also means that we have to be careful that our interventions — say, mandatory mask use — don’t increase the variance of R_t even if they lower its mean.]
Take yourselves back then to early March. In your generic country, the virus spreading in Hubei province, Italy and Iran, and you have a few case counts (less than one hundred). You can see that you are basically a month or so (if you are lucky) behind countries that have locked down. When do you pull the trigger and do the same? Here is what I wrote:
As I write this, it is hard to imagine the information governments were expecting to receive that would have caused them not to act on some type of social distancing. If there was hope, it was not articulated nor in the data. Thus, we are left to speculate. My speculation is that waiting was driven by receiving political news rather than scientific or economic news. Governments may have decided not to shut down if they found that a large proportion of the population would resist those efforts. In many cases, this is why governments implored citizens to engage in social distancing in the hope that they could achieve a reduction in R0 without stronger measures. Those stronger measures included legal requirements to stay at home, which could potentially then be enforced with penalties associated with violations.
In other words, scientifically the decision was obvious. But a politician knew that there were other problems; problems of persuasion. Perhaps most insidious of all was the notion that if they actually pulled the trigger early and succeeded in avoiding the worst of the pandemic, then they may actually be blamed for the economic harm caused by that move. Their opponents would claim the shutdown was not warranted and “the cure was worse than the disease.” Sound familiar? That is pretty much what happened and in countries like Australia, the consequence of that is that they had to experience the virus in a second wave.
The problem is that this was a decision made under uncertainty. And decisions like this tend to end up in what Annie Duke calls, resulting. Resulting occurs when decisions are evaluated by the outcome that resulted. With pandemics, you can’t win. Wait and the virus is out of control and you are blamed for not taking action earlier. Take early action and you are blamed for much ado about nothing. In reality, there was uncertainty and what we need to do is evaluate the decision-making process and not the result.
As Debora MacKenzie notes, when the H1N1 (or swine flu) emerged in 2009, the alert system did its job and some costly actions were taken. The outbreak turned out to be mild but this was, in part, because those born before 1957 were actually immune from the very similar virus that had been a pandemic in 1918–21 and had been circulating since. (Apparently, if you were born between 1957 and 1968 then if this virus comes back you will still be fine.) But following 2009, the WHO was taken to task for declaring a pandemic too soon (or needlessly). Perhaps this is why they seemed to drag their heels when declaring Covid-19 a pandemic. They did so on March 11, 2020. Late or not, when they did, it was a big deal.
Information Reporting
MacKenzie is at her most compelling when she delves back into the history of decisions of past outbreaks over the last two decades. These are stories you wouldn’t have know unless you were paying a lot of attention or were in the midst of a country hit by SARS or MERS. I know I wasn’t and I wish I had paid attention sooner.
But I want to focus on one set of incidents regarding how potential pandemics are reported. There are actually lots of issues here and lots of people who play a role. My main thesis regarding pandemics is that they are fundamentally information problems. Thus, how information is handled is, I would argue, the biggest priority. But during the initial phases of the SARS outbreak in 2003, information reporting was … interesting.
In late 2002, a Canadian alert system picked up pneumonia of unknown origin in China. The WHO only asked the Chinese government about it on February 11th, 2003. It only issued an alert on March 12th. By that time, it had spread to several countries beyond China.
This looks like it is going to be a bad recollection for the WHO. But MacKenzie describes the real narrative. At the time, the WHO could only ask the Chinese government about a potential outbreak if it had received information from them. The reason they only asked China about it was that the Hong Kong health department sent a warning on February 11th. In other words, it wasn’t sluggishness, but the rules that prevented the WHO investigation.
But wait, there’s more. Why did it take another month for an alert to be issued? Well, again it was the rules. The WHO could not issue an alert for a particular country unless the country itself had given permission. At that time, China was still claiming the outbreak was chlamydia. The virus had then spread to Hong Kong. Still part of China and still no permission. It was only when the virus went, via one floor of a Hong Kong hotel, to Canada and Singapore that the WHO could act, which it did. China then requested help and a WHO team arrived the next day. After this, a frustrated WHO broke its own rules to push for further action.
In the end, SARS was contained but not without considerable panic as well as costs associated with delay and a lack of information. I have done no justice to the full story in the summary here. Read the book.
This experience did lead to changes. Critically, the WHO can now ask a government about a virus even if they did not receive information from that country. Also, the WHO can talk about the virus publicly if it is already public (say, because scientists are talking about it in online forums). But the WHO still cannot investigate an outbreak unless invited. Hence, MacKenzie concludes that we are unlikely to really be able to prevent global pandemics without some form of supra-national power. These exist for other treaties such as those to deal with chemical weapons. How this works is interesting:
The CWC’s [Chemical Weapons Convention] real innovation is that someone can charge a member state with not declaring a chemical weapon, or using one illegally, and ask for a surprise “challenge” inspection. Treaty countries have all agreed to “anytime, anywhere” inspections with no right of refusal, except the US, which passed a law allowing it to refuse. No one has ever demanded a challenge inspection, although the OPCW’s destruction of Syrian chemical weapons in 2013–2014 was one in all but name. In another kind of check on bad behavior, the 1987 Montreal Protocol to the treaty banning chemicals that destroy the earth’s protective ozone layer allows member states to slap trade sanctions on countries that break it. It never has, but at one time at least, we all agreed such a threat was appropriate.
I read this and I see an opportunity for economists to play a role in institution design. The reason why these rules are what they are is that governments were worried about ‘bad actors’ coopting the WHO with false information and causing disruptions that may come from that. The rules, therefore, restricted the WHO initially from doing anything without the permission of the government concerned. But they have their own interests too. So there are conflicts all over the shop. Trying to elicit truth in environments such as this is something economists have worked on for half a century. It would be interesting to see them applied here.
The Blame Game Redux
The blame game has consequences. It sets future incentives which is its point. But you had better be sure you are setting the right incentives. It is going to be very unclear that anyone needs sanction or punishment over the initial handling of Covid-19. (Well maybe bats but that is a topic for another time.) That, of course, doesn’t preclude us from casting blame as it became clearer what the right path was. My guess is that sorting this all out will take a couple of decades of work.
