Wei Dai now writes,

The world is probably going to lose 5 2.5-10% of its population (380 190-760 million, see here for the reason for my edit), worse than the Spanish flu even on a percentage basis.

I have some questions about this conclusion and I'm still unsure whether to treat this sort of argument very seriously. But suppose it does come out that between 2.5% to 10% of the world population will die. What are likely results?

I have speculated some of the following. Let me know whether anyone here would agree with my tentative reactions. If the death rate is towards the upper end of 10%, my reactions are probably conservative as I see worse things also happening. Individually I expect each to happen with >50% probability conditional on the above death rate:

  • Net negative GDP growth for 2020, with a 5% or more contraction, making this pandemic much worse than any financial crisis in recent memory (including the 2008 financial crisis).
  • Scientific progress is slowed down, and some technologies take a few more years to come out than they otherwise would have. These include artificial intelligence, plant based meats, anti-aging medicine. This is good if you think that the serial time for AI safety must be lengthened, but bad if
    • you thought that the longer we persisted in our current high-risk period of history, the lower chance we have of surviving
    • you worry about value drift from current norms about AI safety, or society more generally
    • you personally wanted to live until some technological achievement.
  • Life expectancy will drop into the mid 70s in developed countries.
  • Most people will know several people who died of the coronavirus, especially if you knew a lot of old people.
  • A global day of remembrance is declared, and the history books have an entire section on this year as being pivotal in some aspects of world history.
  • The funding for the CDC is multiplied by at least a factor of five.
  • Comprehensive bipartisan legislation is passed that employs tens of thousands of people to continuously monitor the threat of viruses of this magnitude in the future.
  • Lasting injury from the coronavirus is a major disability in the world, and is viewed somewhat akin to how people view polio victims even after polio leaves a community.
  • More young students (>100% increase) want to study epidemiology because they are told that it's a major way to help people in the future, similar to how people are currently being directed into environmental fields as a way to help with climate change.
  • Questions about whether we are taking a current outbreak "seriously" will be the main thoughts on the mind of nearly everyone, including TV anchors, continuing for decades after the event. Here I imagine if a virus sprang up again in some part of the world, world leaders would immediately declare emergency and say that they don't want it to "Turn into 2020" or something like that.
  • Lasting outcry from people who believed that regulation or poor leadership slowed the development of a vaccine or antiviral spawns a political movement to speed up such things in times of crisis.
  • By November, virtually all political discussion is about the coronavirus, and relatively little airtime is given to other causes. Biorisk remains a major issue in the next few presidential elections.
  • Intense criticism of the Chinese regime first for censoring people who spoke about the virus, and then for offering the world optimism when there was none, will be a critical issue, leading to riots and perhaps attempted revolutions.
  • A one time significant political realignment towards people who are young.

Can you come up with anything else?

New to LessWrong?

New Answer
New Comment

3 Answers sorted by


Mar 03, 2020

  • Significant political regime disruption in some places, Iran specifically (probably for the better there)
  • Depending on how much Chinese government fudges with the current numbers they will appear to a) have bungled at first b) actually efficiently handled the problem at an enormous scale and US will appear to do things in the opposite order. Which leads me to...
  • Trump is weakened and Bloomberg's position is going to look very strong: some actual history of crisis management, strong-man people feel comfortable with in times of uncertainty, plus he "owns" one of coutry's top med schools
  • Massive disruption in service sector as people cut down on non-necessities...


Mar 01, 2020


Given who dies from COVID-19, if the scenario suggested plays out, societies will old populations will be hit harder than those with younger populations.

In developed countries this will produce a wealth transfer to the younger via the inheritance they receive (larger than otherwise because less has been spent?)

Not sure that life expectancy really declines even if the numbers for the year blip down.

Probably some themes we could trace out based on both the change in average population age and per capita wealth based on assuming about impact to global trade dynamics.

In developed countries this will produce a wealth transfer to the younger via the inheritance they receive (larger than otherwise because less has been spent?)

It has come to my attention that although the immediate healthcare costs would be dire and extreme, the long-term economic costs might be lower in future years. This is because if something kills a bunch of old people within a short period under tight triage, this may cost the country less money than having to treat the chronic diseases of old age. Furthermore, very old people generally contribute no... (read more)

3 comments, sorted by Click to highlight new comments since: Today at 3:56 AM

You still haven't said why a few days ago you updated to only 5% chance of >50 million global deaths within 1 year. I'm really curious about that.

I'm not sure how much sense this will make, but my meta-credences still say that anything from 2% to 75% chance of disaster are reasonable. I am very uncertain on some very fundamental questions about this virus.

My thoughts are best described as a bunch of scattered intuitions that I have picked up by reading Metaculus comments and analyzing the history of previous pandemics. My estimates are also fluctuating rapidly, and I am worried that if I am constantly spewing out new estimates people will confuse this proper Bayesian updating with ill-judgement. I will keep you updated if or when I write a full inquiry.