Covid-19 Points of Leverage, Travel Bans and Eradication

Covid-19 has be­come a ma­jor topic for dis­cus­sion with talk about many differ­ent in­ter­ven­tions and ideas that might help, from 3-D print­ing parts for res­pi­ra­tors to draft­ing med­i­cal stu­dents into hos­pi­tals to thor­ough hand-wash­ing pro­ce­dures.

How­ever as ra­tio­nal­ists we should be ask­ing which ac­tions have the high­est ex­pected util­ity, not which ac­tions have some pos­i­tive util­ity. In an ex­po­nen­tially grow­ing pro­cess, the ac­tions with the high­est ex­pected util­ity are those ac­tions which in­ter­vene early in the pro­cess, and ac­tions like draft­ing med­i­cal stu­dents which in­ter­vene late in the pro­cess when the dis­ease has already grown to a huge size are “nice to have” but by that point most of the dam­age has been done.

Proper and Prompt Travel Bans do Work

As early as Jan­uary 26th, I called for can­cel­la­tion of flights to limit the spread of covid-19; there was some push­back based on the idea that travel re­stric­tions don’t work which upon closer ex­am­i­na­tion was ac­tu­ally the idea that late or half-hearted travel re­stric­tions don’t work:

Dur­ing the height of the SARS out­break in 2003, he had a col­league who wanted to re­turn to the UK from Toronto, one of the cities most af­fected by the virus. So she caught a do­mes­tic flight from Toronto to Van­cou­ver, then boarded a flight to Lon­don. “When she ar­rived at Heathrow [air­port] and au­thor­i­ties asked her, ‘Have you been to Toronto,’ she said no and walked right through.”

A policy that al­lows peo­ple to travel from an in­fected area to to an un­in­fected area is not a travel ban. It’s con­tain­ment the­ater. A real travel ban would be ground­ing all in­ter­na­tional flights and stop­ping pas­sen­ger trains and boats un­til the dis­ease had been erad­i­cated or at least very well con­tained, as well as ag­gres­sively track­ing down and con­tact trac­ing peo­ple who slipped through be­fore the lock­down, for ex­am­ple us­ing cel­l­phone data from in­tel­li­gence agen­cies. A key point here is that mop­ping up a small num­ber of cases that slip through is in fact pos­si­ble.

It would have been ex­pen­sive to do all this, but the cost of not do­ing it is that the de­vel­oped world is now on lock­down, the stock mar­kets have fallen by around 33% and we have about 10,000 deaths at the time of writ­ing. And we have ended up im­ple­ment­ing the travel bans any­way!


The op­ti­mal strat­egy to defeat the dis­ease is cur­rently the sub­ject of much de­bate. Sev­eral strate­gies have emerged, and a pop­u­lar meme right now is #flat­ten­the­curve. The idea of flat­ten­ing the curve is that if we in­crease the du­ra­tion of the pan­demic, the num­ber of peo­ple in­fected at any one time will be lower and our abil­ity to treat peo­ple prop­erly will be in­creased. Peo­ple put a lot of time into cre­at­ing con­vinc­ing memes and di­a­grams show­ing how this works:

Un­for­tu­nately peo­ple didn’t put much effort into get­ting the num­bers right. Every sin­gle one of these di­a­grams is a steam­ing pile of non­sense be­cause the line for “Health­care Sys­tem Ca­pac­ity” is about 20-50 times too high, which was first pointed out by Joshua Bach. That tiny red line right next to the x-axis is our health sys­tem ca­pac­ity:

(taken from The Im­pe­rial Col­lege COVID-19 Re­sponse Team’s lat­est re­port ).

The UK gov­ern­ment’s “herd im­mu­nity” strat­egy was an­other pos­si­ble way for­ward, but the gov­ern­ment re­versed course on this when they re­al­ized it would in­volve at least a few hun­dred thou­sand deaths.

Con­tain and Eradicate

In my opinion, the cor­rect strat­egy to beat covid-19 whilst min­i­miz­ing losses from this point for­ward is a con­tain-and-erad­i­cate strat­egy. The New England Com­plex Sys­tems In­sti­tute’s writeup on this, writ­ten by Nas­sim Ni­cholas Taleb of Black Swan fame out­lines the strat­egy:

Since lock­downs re­sult in ex­po­nen­tially de­creas­ing num­bers of cases, a com­par­a­tively short amount of time can be suffi­cient to achieve pathogen ex­tinc­tion, af­ter which re­lax­ing re­stric­tions can be done with­out re­sur­gence. …
Fi­nally, the use of ge­o­graphic bound­aries and travel re­stric­tions al­lows for effec­tive and com­par­a­tively low cost im­po­si­tion and re­lax­ation of in­ter­ven­tions. Such a multi-scale ap­proach ac­cel­er­ates re­sponse efforts, re­duces so­cial im­pacts, al­lows for re­lax­ing re­stric­tions in ar­eas ear­lier that are less af­fected, en­ables un­in­fected ar­eas to as­sist in re­sponse in the ares that are in­fected, and is a much more prac­ti­cal and effec­tive way to stop oth­er­wise dev­as­tat­ing out­breaks. …
A few other is­sues are of im­por­tance: They ig­nore the pos­si­bil­ity of su­per­spreader events in gath­er­ings by not in­clud­ing the fat tail dis­tri­bu­tion of con­ta­gion in their model. This leads them to deny the im­por­tance of ban­ning them, which has been shown to be in­cor­rect, in­clud­ing in South Korea. Cut­ting the fat tail of the in­fec­tion dis­tri­bu­tion is crit­i­cal to re­duc­ing R0.


- Close bor­ders and limit in­ter­nal travel, lock­down and hy­giene to drive R0 be­low 1

- Ban large events to cut off the long tail of the R0 distribution

- Use ag­gres­sive test­ing and con­tact trac­ing to clean up any re­main­ing hold­outs, and erad­i­cate the virus on a re­gion-by-re­gion and coun­try-by-coun­try level.

- “Green” re­gions can re­turn to mostly nor­mal life, albeit with­out large events and travel. That means that peo­ple can go back to work and we can re­verse the eco­nomic dam­age.

Con­tain-and-erad­i­cate prob­a­bly re­sults in both less loss of life and less eco­nomic dam­age than any other strat­egy, and we can see this as a con­se­quence of tak­ing an ex­po­nen­tial pro­cess and fight­ing it in the low or­ders of mag­ni­tude rather than the high ones. Flat­ten-The-Curve is bad be­cause a flat curve that lasts for a long time is still, in log-terms, al­most at the max­i­mum power of the virus and there­fore it can do huge amounts of dam­age. Herd-Im­mu­nity and De­liber­ate-In­fec­tion are bad for the same rea­son. The only other sen­si­ble plan I have seen is the idea of rush­ing a vac­cine as quickly as pos­si­ble, but that is be­yond my ex­per­tise.

Travel bans and re­stric­tions dur­ing a pandemic

Why do bor­ders need to be closed now when the virus is already ev­ery­where? Be­cause there is still un­cer­tainty about where the virus is and in what num­bers. The virus wins when peo­ple with differ­ent amounts of virus mix, be­cause ar­eas of high virus can spread to ar­eas of low virus whilst the re­verse pro­cess doesn’t work.

Similarly, if you were cer­tain about who had the virus, this would al­most be triv­ial be­cause all the in­fected could be moved to con­tain­ment fa­cil­ities and ev­ery­one else could get on with run­ning the econ­omy.

The virus wants to max­i­mize en­tropy (virus spread ev­ery­where), hu­man­ity wants to min­i­mize it (all virus in one place), for a given to­tal amount of virus.

The need for bor­ders is a re­sult of this com­bi­na­tion of un­cer­tainty and mix­ing be­ing bad. And as Taleb points out in the NECSI re­view, travel bans and re­stric­tions dur­ing a pan­demic should be mul­ti­level.

As we ap­proach the “endgame” where test­ing is ubiquitous and virus num­bers get closer to 0, bor­ders be­come more im­por­tant, be­cause adding 500 cases to an area with 1 case is much worse than adding 2000 cases to an area with 1000 cases (you have to think in log­a­r­ithms).

Though even when num­bers are high, clos­ing bor­ders is still use­ful and we should still do it; prin­ci­pally be­cause the de­ci­sion to close can lag be­hind rapidly de­vel­op­ing facts on the ground, or worse the de­ci­sion to close bor­ders to a par­tic­u­lar area could leak, at which point peo­ple start ac­tively helping the virus to spread as they flee from the soon-to-be locked down area. Of course no­body would be stupid enough to leak that in­for­ma­tion, right?

What if we do Miti­ga­tion in­stead?

If we do go down the miti­ga­tion path—let­ting most peo­ple get the dis­ease—there are some im­por­tant “dice rolls” that will de­ter­mine how it goes:

  • The rate of long-term com­pli­ca­tions amongst covid-19 sur­vivors,

  • The rate at which young & oth­er­wise healthy peo­ple die when hos­pi­tal treat­ment is de­nied due to overcrowding

  • Whether new phar­ma­ceu­ti­cals like Chloro­quine and Remde­sivir are both effec­tive and scal­able, and how quickly covid-19 evolves re­sis­tance to them

  • Whether sum­mer weather sub­stan­tially slows the spread

  • Whether covid-19 picks up a mu­ta­tion that makes it less lethal, or more lethal

A Test of Rationality

When com­pe­tent Mug­gles make de­ci­sions, they’re usu­ally very em­piri­cal about it. They build a chair with one leg, it falls over, and then they don’t do that again.

Covid-19′s ex­po­nen­tial dy­nam­ics, asymp­tomatic car­ri­ers and long lag time be­tween in­fec­tion and death pun­ished the try-it-and-see ap­proach very hard.

By the time it be­came ab­solutely ob­vi­ous to peo­ple who build one-legged chairs that this was a big deal and needed at­ten­tion, the virus had in­creased both its num­bers and dis­tri­bu­tion most of the way to its goal of in­fect­ing ev­ery hu­man be­ing on the planet.

Covid-19 was a ra­tio­nal­ity test as well as a com­pe­tence test. China failed on ra­tio­nal­ity but passed on com­pe­tence. The West failed hard on ra­tio­nal­ity and is on course for a F+ on com­pe­tence as well. Vox and the other main­stream me­dia who ei­ther mocked those who took it se­ri­ously early, or got on a soap­box talk­ing about racism (which is bad, but was not even re­motely the most im­por­tant thing at that time) should take a rep­u­ta­tional hit. The var­i­ous gov­ern­ment agen­cies that dithered through­out Fe­bru­ary should be in­ves­ti­gated, par­tic­u­larly in the USA.