#18: Coronavirus

Time to apply the precautionary principle.

Mar 4, 2020

I’m an author, organizational sociologist, strategy professor, unsuccessful furniture maker, and Xoogler—this is yet another of my attempts to make sense of the state of not-knowing.

I was in Philadelphia over the weekend talking about what makes an industry flip into one obsessed with continual innovation (not always a good thing) and what leadership should look like in an increasingly uncertain restaurant industry.

Those who work in restaurants have a better visceral understanding of how uncertainty is different from risk (I explain why in the book). But most other people don’t really get it. Nor do they understand how important it is to accurately understand which type of situation you’re facing.

(For a lightweight primer on risk vs uncertainty, you could take a look at issue #1.)

Understanding whether a situation is uncertain or risky is vital because—my long-standing thesis—the world is becoming increasingly uncertain.

The reasons for this are legion, but the most important ones are:

The world system now consists of many increasingly complex and highly optimized subsystems—everything from global banking to supply chains, from immigration to the exchange of data.
These subsystems are becoming more interconnected, and
These subsystems are becoming more interdependent.

(3) means that something which destabilizes one subsystem is more likely to destabilize other subsystems. (2) means destabilization is more likely to spread quickly through the connections between subsystems. (1) means subsystems are more likely to be destabilized by a perturbation. In turn, this means that accurately predicting the behavior of the world system or any of its subsystems becomes nearly impossible.

Understanding that we’re confronting growing uncertainty is not an academic conceit. You’re likely to be really fucked if you respond to a truly uncertain situation by adopting a risk-management approach.

Over dinner on Monday, conversation turned, as it increasingly does, to the coronavirus. (The ship has sailed on calling it 2019-NCoV or even COVID-19.)

The following things seem important:

The asymptomatic infectious period is currently unpredictable in length—most seem to show symptoms in 2 weeks but cases have been reported of 19 and 27 day incubation periods.
The severity of symptoms seems to vary extremely widely among those affected—a small proportion become severely ill and die while the vast majority have mild (or possibly even no) symptoms.
Widespread travel bans have not yet been instituted, and mass gatherings have only been cancelled or prohibited relatively recently.
It may be possible to get reinfected after recovery—though it’s unclear whether there actually was reinfection, and if yes, if it was with the same or a different strain of the coronavirus.
An approved and effective vaccine is unlikely to be widely available within the year.
The global population is immunologically naive to this virus.

Taken together, the only logical conclusion is that a large and still growing population of apparently healthy but infectious people are currently out and about, and have been unknowingly infecting others for an unpredictably long time.

As the world is now extremely well-connected by air travel, these asymptomatic but infectious people are likely to be quite dispersed—and the long asymptomatic period combined with variable symptom severity means they’ll be hard (or even impossible) to detect. If reinfection is possible, that considerably compounds the problem.

The lightweight containment strategies which most affected countries have been using focus on identifying and isolating transmission chains (tracing the chain of infected people backwards). This approach is relatively undisruptive and probably effective in containing infectious disease outbreaks that are more like historic ones have been.

But the specific characteristics of coronavirus mean these lightweight containment strategies probably won’t work. It’s hard to contain a virus by tracing transmission chains when infected people can be contagious for many days before they show symptoms of infection.

China seemed to very quickly go past containment into mitigation. At the moment, these expensive mitigation measures seem to be bearing some fruit as the epidemiological curve in China flattens down. That same curve is trending upwards in the rest of the world—and few other states have the apparatus necessary to nearly completely shut down traffic and commerce in cities of the size (and areas of the extent) that China has.

If containment efforts fail (which seems to be happening) and mitigation measures—which will necessarily include widespread travel bans, mandatory suspension of mass events, school/work shutdowns, and other crushingly expensive and politically/socially unpopular decisions—aren’t put in place very soon, a sizeable proportion of the world’s population could quickly be infected.

Many commentators who brush off pandemic concerns and the idea of introducing widespread mitigation measures as paranoia or overreaction have noted that the flu kills more people every year than coronavirus has since the outbreak began.

This is true, but a specious argument.

The global population is immunologically experienced with the flu, and there is an existing infrastructure for developing, producing, and distributing a validated vaccine for every flu season. Neither is true for coronavirus.

The rate at which infected patients die from coronavirus is low now when they can receive timely and appropriate care. If there is a pandemic, the mortality rate will inevitably rise as healthcare systems reach, then exceed capacity. It is a peak flow problem. Mortality rates, incidentally, are unlikely to rise only linearly with infection rates.

But that’s not the worst of it. The real impact of globally pandemic coronavirus is that it will disrupt global supply chains enough that the disruption will spread to other world subsystems. If this happens, the impact globally will soon far exceed the loss of life directly caused by coronavirus infection.

Disrupt electronics supply chains and cellphones and laptops become more expensive or entirely unavailable—but really this is not so very important. Disrupt medical equipment and pharmaceutical supply chains and patients (with coronavirus or other diseases) die who would otherwise recover, while long-term healthcare costs go up—this is vitally important. Disrupt food supply chains and food prices rise, then severe food shortages appear—with the potential for triggering civil unrest. These supply chains are complex and highly optimized and thus fragile and susceptible to disruption.

More generally, our world economy now depends on continual growth. It will be severely affected by the massive and prolonged economic contraction coronavirus will cause. It is entirely plausible that many countries will enter prolonged recessions.

Despite the enormous potential danger to what might melodramatically be described as world civilization, most national and international health organizations seem to be taking a risk-management approach by choosing lower-cost containment strategies instead of moving quickly to much more expensive, politically unpopular mitigation strategies. Unsurprisingly, they have been weighing benefits against costs when deciding between containment and mitigation, as is appropriate in a risk situation when the range of possible outcomes is clear and thus costs and benefits of different actions are clear as well.

The problem is: This coronavirus epidemic is a situation of true and great uncertainty, not of risk.

While the costs of containment and mitigation are relatively clearly understood, the range of possible outcomes and their relative probabilities is unknown. Coronavirus might destroy the globalized civilization we’ve built up over the last few decades. Given the evidence of the last two days alone, coronavirus is not going to fizzle out like SARS did. How can the benefits of containment and mitigation be compared against each other when their consequences are shrouded in uncertainty?

Put plainly, containment looks cheap compared to mitigation. But conventional cost-benefit analyses don’t—and cannot—address how catastrophic it could be to get the coronavirus response strategy wrong.

Approaching the inherent uncertainty of a potential coronavirus pandemic with a risk-management mindset is likely to be a recipe for disaster.

I’ve been getting emails asking for the newsletter to go biweekly because some people are experiencing New Yorker syndrome. (I think that’s a compliment.) If you have thoughts, please reply to this issue to tell me what you think.

Irresistible.

If you found this useful or thought-provoking, you should definitely share it indiscriminately with masses of people. You can find previous issues of The Uncertainty Mindset here. I’m on the web at www.vaughntan.org, on Twitter @vaughn_tan, Instagram @vaughn.tan, or by email at <uncertaintymindset@vaughntan.org>.