To Save the Economy, Hit the Coronavirus Fast and Hit it Hard

Half measures will cost lives and do more damage than fighting all-out

You may have noticed that we are in the early stages of the most significant pandemic in 100 years. You may also have noticed that the fight against COVID-19 has been escalating rather significantly. But how hard do we really need to fight this coronavirus? Do we really have to shut everything down?

Unfortunately, epidemic models suggest yes. Half measures will be nowhere near enough, and it’s possible they could make the economic fallout even worse.

Flattening the curve

Fighting this virus all comes down to reducing how well it spreads. In epidemic parlance, that’s the viral reproduction rate, or R0. Models suggest that if we did nothing to fight this virus, it would infect most of the world in just a few months. Several percent of the population would die.

If we do nothing, the coronavirus will tear through society like wildfire. Flattening the curve could save some lives, but it will also make the epidemic last longer.

The worst case won’t happen, so long as we all fight back. We’re slowing the viral spread through social distancing, which can save millions of lives and reduce the overload on the health care system. This is the “flattening the curve” strategy you’ve probably seen a lot about.

Unfortunately, pushing down the peak of the curve also spreads it out. That means the outbreak lasts longer, potentially more than a year. With no vaccines or treatments, the only weapons we have to flatten the curve are also socially painful. The longer the virus rampages, the harder those measures will be to sustain, and the more they will hurt the economically vulnerable.

Beating it back quick

So what’s the game plan, if doing nothing kills millions, but flattening the curve makes the epidemic last even longer? This is where R0 comes in again. If we can slow the spread enough, pushing R0 below 1, suddenly the game changes. In fact, the name of the game literally changes — we leave mitigation and begin suppression.

If we slow the spread of the coronavirus by more than 63%, it begins to die out, rather than spreading.

Now each infected person infects less than 1 other person on average, so the virus can no longer spread exponentially. Instead, it begins dropping. For the COVID-19 coronavirus, the magic number is to reduce the viral rate of spread by about 63%.

Importantly, the harder we fight its spread, the faster we defeat the virus. If we slow the spread by 70%, we could drive the new daily cases down to a manageable number and ease restrictions after about 6 months. Slow it by 80%, and it could be 2–3 months. Slow the virus by 90%, and we could largely beat it back it in only 4–8 weeks, like China and South Korea appear to have done (at least for now).

Months until the number of new daily cases drops below a manageable threshold, say 10 or 100. Flattening the curve initially makes the epidemic last longer, unless we cross into suppression. Once we do, the harder we fight, the quicker it’s over. The curves are capped at 18 months, assuming that’s when we’ll have a vaccine.

Minimizing collateral damage

One legitimate concern is that if we overreact, we might cause more collateral damage to our daily lives and economy than we need to. That’s not just being selfish and miserly — economic hardship causes real suffering, and poverty takes lives. But, perhaps counterintuitively, if we’re going to stamp out the virus, it would likely be best to do it as fast as possible.

Assuming we’re not willing to watch a few million bodies pile up outside our hospitals to be removed by the military, we will need strong social distancing restrictions, which will cause a lot of collateral damage.

The amount of damage is hard to predict, but as the most basic assumption, let’s say it’s proportional to how aggressively we reduce viral spread. You might think that means less aggressive measures would do less damage, but remember the damage will keep accumulating as long as we need restrictions. Since more aggressive action will bring the virus under control even quicker, the strongest response may actually do the least amount of collateral social damage.

Assuming we decide not to allow this coronavirus to kill millions, pulling out all the stops to defeat it as quickly as possible may cause less social and economic damage. Even if fighting harder makes the weekly economic damage worse, the outbreak will be over quicker, and less total damage will be done.

Buying time for a vaccine

Of course, even if we beat the virus back, it will still be lurking in the shadows, ready to break out again at any time. However, once we knock it down to few enough cases per day, we stand a fighting chance of containing them with less disruptive measures: contact tracing, testing (fingers crossed), and more targeted quarantines, rather than massive social disruption.

The real end game here is a vaccine. Once we have that, which will take 1–2 years, we can deliver the knockout blow and declare victory over this virus for good. Assuming anti-vaxxers don’t ruin it for us, of course.

Which door: A, B, or C?

So now we have three choices. A, we could not do enough and allow millions of people to die. (Note this will cause massive social upheaval not shown in the orange curve). B, we could take half measures, which would save a portion of those lives, but also prolong the pandemic and severely damage the economy. C, we can go all-out. This will require acute social disruption, but it would prevent nearly all the possible deaths. And once we choose option C, we may as well fight as hard as we can. If we fight this virus as hard as possible, it could be over in a month or two, with less economic collateral damage than if we held back. Just like ripping off a bandaid, this will be least painful if we do it hard and fast.

For a rather longer analysis that comes to a similar conclusion, The Hammer and the Dance is an excellent read if you have the time.

Appendix: Math and assumptions

This analysis is based on an SEIR (susceptible-exposed-infectious-recovered) model published in The Lancet (no affiliation), implemented here as a set of coupled ordinary differential equations. R0 is assumed to be 2.68, the incubation period is 6 days, and the serial interval is 8.4 days. The base case fatality rate is 0.94%, but it triples when hospital capacity is overwhelmed. The hospitalization rate for COVID-19 patients is assumed to be 14%, and the US has 2.9 hospital beds per 1000 people.

The analysis here did not consider travel or other population structure, but models including those factors found largely similar results.

It is possible that summer will help us with some of the job, reducing viral spread through higher heat and humidity, but the fact that southern hemisphere countries are dealing with the same outbreaks shows that we can’t rely on the summer alone.