Issue 70

Further thoughts on the coronavirus

Delivered on 23 March 2020 by Justin Pyvis. About a 7 min read.

Most of the world has, to varying degrees, shut down. That means there's very little happening in the tech sector, outside of SoftBank's desperate attempt to raise another $US10 billion (the definition of throwing good money after bad), or the US senate's determination to quickly pass through laws that undermine people's privacy.

So this week I'm going to write about the coronavirus again.

You may have heard the story of the Northern Italian town of Vò, about 65km west of Venice and sandwiched in the middle of Italy's coronavirus 'redzone'. If not, great, because I'm about to tell you.

Vò had some of Italy's earliest coronavirus cases and was the first city to experience a COVID-19 death, leading Italian researchers to test the entire town - around 3,300 people - for COVID-19. What did they find? That nearly 3% of the population were infected, many of them without symptoms. So they quarantined the entire town as a precaution:

"They were asked not to go out, and not to have contact with any other people," Professor Crisanti said.  The researchers decided against sending patients to hospital to prevent them spreading the disease there. "In principle many people in the hospital were infected. Many doctors, many nurses, many patients. This could be a major source of infection," he said.

After two weeks of quarantine, the researchers carried out another round of mass testing in Vo. The rate of COVID-19 infection had dropped from nearly 3 per cent to 0.41 per cent. Luca Zaia, the governor of Veneto, has now declared the town of Vo the "healthiest place in Italy".

Professor Crisanti has one regret:

"As soon as we made data available that the infection rate in Vo was 3 per cent, they should have rung the alarm bell," he says. "If at that time we had locked down all the cities with cases and carried out extensive testing, we would not be in this situation."

I'd say he has a good point - by the time this issue has been distributed, Italy will have recorded its 5,000th COVID-19 related death from well over 50,000 reported cases. The country is now in a full blown health and economic crisis.

There are some important lessons we can take away from the experience in Vò.

1. Think exponentially, not linearly

No politician wants to act prematurely and risk being accused of overreacting. They would much rather wait until the situation worsens, declare themselves a "wartime President", and re-frame the narrative to paint themselves as the hero when things eventually get better (and they will, provided governments don't cause too much damage in the meantime - please, no price controls!).

Bureaucrats face similar incentives. It would be a career limiting move to risk prematurely raising the alarm, especially if senior politicians are still telling people to "just get about their daily business". No one who knows how to get themselves promoted in a bureaucratic organisation - and especially not someone in human resources - will risk challenging the prevailing rules and advice unless it first comes from the top-down (i.e. a politician), which means by the time appropriate action is taken, it's already too late.

Unfortunately, those incentives lead to precisely the opposite behaviour to what is required to stem a viral outbreak, because viruses spread exponentially, not linearly. That leads people to underestimate the risks in the early phase when case numbers are low and appear to be growing linearly, and overestimate the risks in the latter phase (epidemics generally follow a bell curve pattern - the red line below).

2. Early social distancing is key, even before you think you need to

When I say early, I mean early. Exponential growth and the inevitable lags before a case is confirmed - the average case shows symptoms only after about 5 days, and people may not report for testing until much later than that (if at all) - means if you only watch the case data you will always be at least a week behind reality.

Let's put that into context. If the confirmed case data are a week behind, you have 1,000 cases, and cases are growing at least 20% per day (basically Australia at the time of writing), that means you probably already have around 3,500 cases. But by the time the 3,500th case is confirmed, you're actually at about 13,000 cases. And so on.

See the problem for policy-makers looking solely at the number of confirmed cases? Everything they do, when they eventually get around to doing it, will be at least a week too late.

What should be done instead is as soon as community transmission is confirmed - even just a handful of cases - the first, relatively low-cost social distancing measures should be implemented. Given the exponential nature of viruses, even a small reduction in the rate of infections can dramatically slow the spread.

Such measures include instructing office workers to work from home or in much smaller, segregated teams. Large events should be postponed where possible, or cancelled. Mandatory quarantines for all arrivals and social distancing should be strongly incentivised, with both carrot and stick. The floating Petri dishes known as cruise liners should be suspended immediately.

Those small acts alone would delay the spread, flatten the curve and possibly reduce the need for heavy-handed restrictions and the associated economic costs further down the road.

3. Test early, test often, and test widely.

Most people won't grasp the magnitude of the outbreak from the confirmed case data alone, at least not early enough for it to matter. So you need to test, test and test some more, and not just people with symptoms.

Given the incentives they face, unless policy-makers are able to comprehend the extent of the outbreak in the entire population, they won't be able to act appropriately. That means representative samples of the population should be taken and tested as often as possible, otherwise outbreaks will grow unnoticed to the point that localised lock-downs, as done in Vò, become infeasible.

If localised isolation won't work because you can't find outbreaks early enough, you're left with no option but relatively high-cost, mass isolation - a move that effectively resets the clock, allowing you to rapidly slow the spread before reverting to a localised lock-down model as outbreaks appear again (and they will). Unfortunately, that's the only option left for the majority of the world's governments, which have by and large botched their responses to COVID-19.

Finally, remember that even if measures to slow the virus are working, the data will look worse before it gets better. Given the lags mentioned above, it will be at least a week after any measure is implemented before dividends are seen in the data.

Stay safe and enjoy the rest of this week's issue. Cheers,

— Justin

Other bits of interest

Debt is a magnifier

John Cochrane writes:

Our economy runs on a river of debt, and the debt clock keeps ticking when the economy shuts down. Had everyone kept a few months of cash around, things would be fine. But many did not. Now we are seeing the beginnings of a scramble for cash, as people and businesses try to sell assets or borrow.

Money troubles spread like a virus. When a business cannot pay, its creditors, employees, investors and banks are in trouble. And if people worry that banks and other institutions are going to fail, they run to get money out—and we have a crisis.

Many companies - airlines in particular - were taking advantage of "unconventional" monetary policy to borrow a lot to buy back their own stock. That's fine when the going's good, and there's nothing wrong with buy backs in and of themselves (whether the tax system should incentivise debt and buybacks, as it currently does, is a separate question). But it also leaves you incredibly exposed when cash flows dry up. As John concludes, "When it's over, we need to start taking seriously that if you're too big to fail, you're too big to borrow. Airlines, this means you".


The world is working from home...

...and it will have a lasting effect on the workplace well beyond the coronavirus.

For jobs where it's possible, everyone should already be working from home, regardless of your country's case numbers. That's great for streaming services, online delivery services, Zoom, Slack and Microsoft Teams (but you should really be using Keybase or, if your workplace gives you the choice), but not so great for WeWork, airlines and corporate travel agents, or Europe's net neutrality.


The post-coronavirus world is at risk

There is a large risk that many countries turn more protectionist and nativist in the aftermath of COVID-19. That would be a great tragedy, and it will take a lot of work to get governments to roll-back their temporary increases in power to combat the coronavirus. Just as governments were almost universally too slow to act to stop the virus spreading, they will be too slow to wind back their restrictions on activity.


Issue 70: Further thoughts on the coronavirus was compiled by Justin Pyvis and delivered on 23 March 2020.