Hide table of contents

The views expressed here are my own, not those of my employers.

Summary

  • I previously estimated an astronomically low annual probability of a conflict causing human extinction fitting Pareto distributions to increasingly rightmost sections of the tail distribution of the annual conflict deaths as a fraction of the global population. In this post, I run a similar analysis for annual epidemic/pandemic deaths.
  • The expected damage from epidemics/pandemics is very concentrated in the most severe ones, with 94.0 % of the annual epidemics/pandemics deaths as a fraction of the global population coming from ones whose such fraction is at least 0.1 %.
  • There has been a downward trend in the logarithm of the annual epidemic/pandemic deaths as a fraction of the global population, with the coefficient of determination (R^2) of the linear regression of it on the year being 38.5 %.
  • In contrast to the conclusion of my past analyses on terrorist attacks, wars and conflicts, a pandemic can possibly cause human extinction, at least on priors. The tail distribution of the annual epidemic/pandemic deaths as a fraction of the global population does not decay fast enough for the annual probability of a pandemic causing human extinction to become astronomically low.

Introduction

I previously estimated an astronomically low annual probability of a conflict causing human extinction fitting Pareto distributions to increasingly rightmost sections of the tail distribution of the annual conflict deaths as a fraction of the global population. In this post, I run a similar analysis for annual epidemic/pandemic deaths.

I focussed on Pareto distributions for simplicity, and because these are arguably the most commonly used to model tail risk. One could argue they have overly thin tails due the possibility of dragon kings, i.e. flatter sections of the right tail (in the unobserved domain), but I assume steep sections cannot be ruled out either. Consequently, I supposed Pareto distributions still offer a good prior.

Methods

I got the epidemic/pandemic deaths from 1500 to 2023 based on data from Marani et. al 2021 on the start and end years as well as deaths by epidemic/pandemic[1]. I stipulated the deaths of each epidemic/pandemic are distributed uniformly from its start to end year. 

I assumed Marani et. al’s data underestimate the epidemic/pandemic deaths due to underreporting.

  • I got no epidemic/pandemic deaths with the procedure above in 8 of 524 (= 2023 - 1500 + 1) years, i.e. 1.53 % (= 8/524) of them. However, I wanted the deaths to be positive in all years to fairly assess the linear regression of the logarithm of the annual epidemic/pandemic deaths as a fraction of the global population on the year. So I supposed the epidemic/pandemic deaths in each of the aforementioned 8 years to be 597, which is half of the ratio between 5 k deaths (lowest value outside these 8 years; see previous footnote) and the mean epidemic/pandemic duration of 4.19 years.
  • For the other years, I considered Marani et. al’s annual epidemic/pandemic deaths as a fraction of the actual annual epidemic/pandemic deaths are represented by a piecewise linear function of the year[2]. I asked Marco Marani on July 2 about some points to define it, but I have not heard back[3]. So I used the function of my analysis of conflict deaths, which is defined by the following points (in agreement with guesses from Peter Brecke, who built the Conflict Catalog dataset):

    • In 1400, 4 %.

    • In 1700, 10 %.

    • In 1800, 22.5 %.

    • In 1900, 90 %.

    • In 2000, 100 %.

      • I also relied on a value of 100 % for 2023, as the above implies underreporting decreases with time.

I fit Pareto distributions to the 2, 3, … and 524 (= 2023 - 1500 + 1) years with the most epidemic/pandemic deaths as a fraction of the global population. To do this:

  • I get the slope and intercept of linear regressions of the logarithm of (sections of) the tail distribution on the logarithm of the annual epidemic/pandemic deaths as a fraction of the global population.
  • Since the tail distribution of a Pareto distribution is P(X > x) = (“minimum”/x)^“tail index”, ln(P(X > x)) = “tail index”*ln(“minimum”) - “tail index”*ln(x) = “intercept” + “slope”*ln(x), so I determine the parameters of the Pareto distributions from:
    • “Tail index” = -“slope”.
    • “Minimum” = e^(“intercept”/“tail index”).

Then I obtain the annual probability of a pandemic causing human extinction from that of conflict deaths exceeding the global population, which is P(X > 1) = “minimum”^“tail index”. This decreases with the tail index, given the minimum of each Pareto distribution is lower than 1.

The calculations are in this Sheet.

Results

Historical annual epidemic/pandemic deaths as a fraction of the global population

Preprocessing

Basic stats

StatisticAnnual epidemic/pandemic deaths as a fraction of the global population
Mean0.236 %
Minimum0
5th percentile1.19*10^-6
10th percentile3.60*10^-6
Median0.0276 %
90th percentile0.414 %
95th percentile0.684 %
Maximum10.3 %

Worst years

NNth highest annual epidemic/pandemic deaths as a fraction of the global populationYearLocation of the epidemics/pandemics
110.3 %1520Americas, Dominican Republic, Haiti, Ireland and Mexico
210.1 %1519China, Dominican Republic, Haiti, Ireland and Mexico
37.39 %1545Americas, China and India
47.34 %1546Americas
57.32 %1547Americas
67.30 %1548Americas
72.02 %1576Americas, England, France, Ireland and Italy
82.01 %1577Americas, England, Ireland and Italy
91.96 %1578Americas and France
101.45 %1770England, Germany, India, Iraq, Ireland, Italy, Russia, Scotland and Sweden

Linear regression of the logarithm of the annual epidemic/pandemic deaths as a fraction of the global population on the year

Slope (1/year)InterceptCoefficient of determination
-0.010610.238.5 %

Risk by severity

Annual epidemic/pandemic deaths as a fraction of the global populationYearsYears as a fraction of the total

Conditional annual epidemic/pandemic deaths as a fraction of the global population

Unconditional annual epidemic/pandemic deaths as a fraction of the global population

Unconditional annual epidemic/pandemic deaths as a fraction of the global population
MinimumMaximum
0Infinity524100 %

0.236 %

0.236 %

100 %
010^-661.15 %

9.05*10^-7

1.04*10^-8

4.38*10^-6
10^-60.001 %7414.1 %

3.24*10^-6

4.57*10^-7

0.0193 %
0.001 %0.01 %8115.5 %

0.00471 %

0.00073 %

0.308 %
0.01 %0.1 %20238.5 %

0.0342 %

0.0132 %

5.58 %
0.1 %1 %14728.1 %

0.368 %

0.103 %

43.6 %
1 %10 %122.29 %

3.50 %

0.0802 %

33.9 %
10 %100 %20.382 %

10.2 %

0.0390 %

16.5 %

Tail distribution

Pandemics tail risk

The tail distribution decays faster for higher tail indices. A tail index of 1 (10) means annual epidemic/pandemic deaths as a fraction of the global population are 10 % (10^-10) likely to become 10 times as large.

Discussion

Historical annual epidemic/pandemic deaths as a fraction of the global population

The expected damage from epidemics/pandemics is very concentrated in the most severe ones, with 94.0 % (= 0.436 + 0.339 + 0.165) of the annual epidemics/pandemics deaths as a fraction of the global population coming from ones whose such fraction is at least 0.1 % (see last table).

The highest annual epidemic/pandemic deaths as a fraction of the global population were 10.3 % in 1520. This involved pandemics/epidemics in the Americas, Dominican Republic, Haiti, Ireland and Mexico.

There has been a downward trend in the logarithm of the annual epidemic/pandemic deaths as a fraction of the global population, with the R^2 of the linear regression of it on the year being 38.5 %. I guess the sign of the slope is resilient against changes to my modelling of the underreporting. One may argue the aforementioned logarithm will increase in the next few decades based on inside view factors such as technology becoming cheaper and more powerful. Nevertheless, technology also became cheaper and more powerful during the period of 1500 to 2023 covered by my data, but these still suggest a decreasing logarithm of the annual epidemic/pandemic deaths as a fraction of the global population.

Pandemics tail risk

Increasing the lowest annual epidemic/pandemic deaths as a fraction of the global population included in the linear regression of the tail distribution, i.e. relying on increasingly rightmost sections of the tail, the annual probability of a pandemic causing human extinction (see last graph):

  • Decreases to 0.0206 % for 0.202 % lowest annual epidemic/pandemic deaths as a fraction of the global population.
  • Then increases to 0.0823 % for 1.31 % lowest annual epidemic/pandemic deaths as a fraction of the global population.
  • Then decreases to 9.76*10^-7 for 7.30 % lowest annual epidemic/pandemic deaths as a fraction of the global population.
  • Then increases to 0.00107 % for 7.39 % lowest annual epidemic/pandemic deaths as a fraction of the global population.
  • Then decreases to 1.46*10^-34 for 10.1 % lowest annual epidemic/pandemic deaths as a fraction of the global population.

The median superforecaster and domain expert of The Existential Risk Persuasion Tournament (XPT) predicted an annual probability of an engineered pathogen causing human extinction in the 78 years (= 2100 - 2023 + 1) from 2023 to 2100 of 1.28*10^-6 (= 1 - (1 - 10^-4)^(1/78)) and 0.0129 %[4] (= 1 - (1 - 10^-4)^(1/78)).

My only astronomically low annual probability of a pandemic causing human extinction of 1.46*10^-34 relies on just the 2 rightmost points of the tail. I believe these should get the most weight to predict extinction risk, but that such a tiny sample size is clearly insufficient to conclude the prior annual probability of a pandemic causing human extinction is astronomically low. This is in contrast to the conclusion of my past analyses on terrorist attacks, wars and conflicts.

The tail distribution still has to eventually decay abruptly for annual epidemic/pandemic deaths as a fraction of the global population that are higher than the ones I studied, as deaths are limited to the global population. However, such an abrupt decrease is not born out of past data. So, at least on priors, a pandemic can possibly cause human extinction.

  1. ^

     I assumed 5 k deaths (= (0 + 10)/2*10^3) for epidemics/pandemics qualitatively inferred (said) to have caused less than 10 k deaths, which are coded as having caused -999 (0) deaths. I also considered the deaths from COVID-19, which is not in the original dataset.

  2. ^

     By Marani et. al’s annual epidemic/pandemic deaths, I mean those I got based on Marani et. al’s data, but Marani et. al 2021 do not (explicitly) provide data on the annual epidemic/pandemic deaths.

  3. ^

     Marco is the 1st author of Marani et. al 2021.

  1. ^

     I calculated these values annualising the ones in Table 3.

Show all footnotes
Comments4


Sorted by Click to highlight new comments since:

Could you please expand on why you think a Pareto distribution is appropriate here? Tail probabilities are often quite sensitive to the assumptions here, and it can be tricky to determine if something is truly power-law distributed.

When I looked at the same dataset, albeit processing the data quite differently, I found that a truncated or cutoff power-law appeared to be a good fit. This gives a much lower value for extreme probabilities using the best-fit parameters. In particular, there were too few of the most severe pandemics in the dataset (COVID-19 and 1918 influenza) otherwise; this issue is visible in fig 1 of Marani et al. Could you please add the data to your tail distribution plot to assess how good a fit it is?

A final note, I think you're calculating the probability of extinction in a single year but the worst pandemics historically have lasted multiple years. The total death toll from the pandemic is perhaps the quantity most of interest.

Thanks for the relevant points, Joshua. I strongly upvoted your comment.

Could you please expand on why you think a Pareto distribution is appropriate here?

I did not mean to suggest a Pareto distribution is appropriate, just that it is worth considering.

Tail probabilities are often quite sensitive to the assumptions here, and it can be tricky to determine if something is truly power-law distributed.

Agreed. In my analysis of conflict deaths, for the method where I used fitter:

The 5th and 95th percentile annual probability of a conflict causing human extinction are 0 and 5.02 % [depending on the distribution]


When I looked at the same dataset, albeit processing the data quite differently, I found that a truncated or cutoff power-law appeared to be a good fit. This gives a much lower value for extreme probabilities using the best-fit parameters. In particular, there were too few of the most severe pandemics in the dataset (COVID-19 and 1918 influenza) otherwise; this issue is visible in fig 1 of Marani et al. Could you please add the data to your tail distribution plot to assess how good a fit it is?

I did not get what you would like me to add to my tail distribution plot. However, I added here the coefficients of determination (R^2) of the regressions I did.

A final note, I think you're calculating the probability of extinction in a single year but the worst pandemics historically have lasted multiple years. The total death toll from the pandemic is perhaps the quantity most of interest.

Focussing on the annual deaths as a fraction of the global population is useful because it being 1 is equivalent to human extinction. In contrast, total epidemic/pandemic deaths as a fraction of the global population in the year in which the epidemic/pandemic started being equal to 1 does not imply human extinction. For example, a pandemic could kill 1 % of the population each year for 100 years, but population remain constant due to births being equal to the pandemic deaths plus other deaths.

However, I agree interventions should be assessed based on standard cost-effectiveness analyses. So I believe the quantity of most interest which could be inferred from my analysis is the expected annual epidemic/pandemic deaths. These would be 2.28 M (= 2.87*10^-4*7.95*10^9) multiplying:

  • My annual epidemic/pandemic deaths as a fraction of the global population based on data from 1900 to 2023. Earlier years are arguably not that informative.
  • The population in 2021.

The above expected death toll would rank as 6th in 2021.

For reference, based on my analysis of conflicts, I get an expected death toll of conflicts based on historical data from 1900 to 2000 (also adjusted for underreporting), and the population in 2021 of 3.83 M (= 2.87*10^-4*7.95*10^9), which would rank above as 5th.

Here is a graph with the top 10 actual causes of death and expected conflict and epidemic/pandemic deaths:

Are the high numbers of deaths in the 1500s old world diseases spreading in the new world? If so, that seems to overestimate natural risk: the world's current population isn't separated from a larger population that has lots of highly human-adapted diseases.

In the other direction, this kind of analysis doesn't capture what I personally see as a larger worry: human-created pandemics. I know you're extrapolating from the past, and it's only very recently that these would even have been possible, but this seems at least worth noting.

Thanks for the comment, Jeff.

Are the high numbers of deaths in the 1500s old world diseases spreading in the new world?

Yes, and deaths are especially high in the 1500s given my assumption of high underreporting then.

If so, that seems to overestimate natural risk: the world's current population isn't separated from a larger population that has lots of highly human-adapted diseases.

Agreed. Personally, I guess the annual probability of a natural pandemic causing human extinction is lower than 10^-10.

In the other direction, this kind of analysis doesn't capture what I personally see as a larger worry: human-created pandemics. I know you're extrapolating from the past, and it's only very recently that these would even have been possible, but this seems at least worth noting.

I think it is interesting that:

There has been a downward trend in the logarithm of the annual epidemic/pandemic deaths as a fraction of the global population, with the R^2 of the linear regression of it on the year being 38.5 %. I guess the sign of the slope is resilient against changes to my modelling of the underreporting. One may argue the aforementioned logarithm will increase in the next few decades based on inside view factors such as technology becoming cheaper and more powerful. Nevertheless, technology also became cheaper and more powerful during the period of 1500 to 2023 covered by my data, but these still suggest a decreasing logarithm of the annual epidemic/pandemic deaths as a fraction of the global population.

Curated and popular this week
 ·  · 23m read
 · 
Or on the types of prioritization, their strengths, pitfalls, and how EA should balance them   The cause prioritization landscape in EA is changing. Prominent groups have shut down, others have been founded, and everyone is trying to figure out how to prepare for AI. This is the first in a series of posts examining the state of cause prioritization and proposing strategies for moving forward.   Executive Summary * Performing prioritization work has been one of the main tasks, and arguably achievements, of EA. * We highlight three types of prioritization: Cause Prioritization, Within-Cause (Intervention) Prioritization, and Cross-Cause (Intervention) Prioritization. * We ask how much of EA prioritization work falls in each of these categories: * Our estimates suggest that, for the organizations we investigated, the current split is 89% within-cause work, 2% cross-cause, and 9% cause prioritization. * We then explore strengths and potential pitfalls of each level: * Cause prioritization offers a big-picture view for identifying pressing problems but can fail to capture the practical nuances that often determine real-world success. * Within-cause prioritization focuses on a narrower set of interventions with deeper more specialised analysis but risks missing higher-impact alternatives elsewhere. * Cross-cause prioritization broadens the scope to find synergies and the potential for greater impact, yet demands complex assumptions and compromises on measurement. * See the Summary Table below to view the considerations. * We encourage reflection and future work on what the best ways of prioritizing are and how EA should allocate resources between the three types. * With this in mind, we outline eight cruxes that sketch what factors could favor some types over others. * We also suggest some potential next steps aimed at refining our approach to prioritization by exploring variance, value of information, tractability, and the
 ·  · 5m read
 · 
[Cross-posted from my Substack here] If you spend time with people trying to change the world, you’ll come to an interesting conundrum: Various advocacy groups reference previous successful social movements as to why their chosen strategy is the most important one. Yet, these groups often follow wildly different strategies from each other to achieve social change. So, which one of them is right? The answer is all of them and none of them. This is because many people use research and historical movements to justify their pre-existing beliefs about how social change happens. Simply, you can find a case study to fit most plausible theories of how social change happens. For example, the groups might say: * Repeated nonviolent disruption is the key to social change, citing the Freedom Riders from the civil rights Movement or Act Up! from the gay rights movement. * Technological progress is what drives improvements in the human condition if you consider the development of the contraceptive pill funded by Katharine McCormick. * Organising and base-building is how change happens, as inspired by Ella Baker, the NAACP or Cesar Chavez from the United Workers Movement. * Insider advocacy is the real secret of social movements – look no further than how influential the Leadership Conference on Civil Rights was in passing the Civil Rights Acts of 1960 & 1964. * Democratic participation is the backbone of social change – just look at how Ireland lifted a ban on abortion via a Citizen’s Assembly. * And so on… To paint this picture, we can see this in action below: Source: Just Stop Oil which focuses on…civil resistance and disruption Source: The Civic Power Fund which focuses on… local organising What do we take away from all this? In my mind, a few key things: 1. Many different approaches have worked in changing the world so we should be humble and not assume we are doing The Most Important Thing 2. The case studies we focus on are likely confirmation bias, where
 ·  · 1m read
 · 
I wanted to share a small but important challenge I've encountered as a student engaging with Effective Altruism from a lower-income country (Nigeria), and invite thoughts or suggestions from the community. Recently, I tried to make a one-time donation to one of the EA-aligned charities listed on the Giving What We Can platform. However, I discovered that I could not donate an amount less than $5. While this might seem like a minor limit for many, for someone like me — a student without a steady income or job, $5 is a significant amount. To provide some context: According to Numbeo, the average monthly income of a Nigerian worker is around $130–$150, and students often rely on even less — sometimes just $20–$50 per month for all expenses. For many students here, having $5 "lying around" isn't common at all; it could represent a week's worth of meals or transportation. I personally want to make small, one-time donations whenever I can, rather than commit to a recurring pledge like the 10% Giving What We Can pledge, which isn't feasible for me right now. I also want to encourage members of my local EA group, who are in similar financial situations, to practice giving through small but meaningful donations. In light of this, I would like to: * Recommend that Giving What We Can (and similar platforms) consider allowing smaller minimum donation amounts to make giving more accessible to students and people in lower-income countries. * Suggest that more organizations be added to the platform, to give donors a wider range of causes they can support with their small contributions. Uncertainties: * Are there alternative platforms or methods that allow very small one-time donations to EA-aligned charities? * Is there a reason behind the $5 minimum that I'm unaware of, and could it be adjusted to be more inclusive? I strongly believe that cultivating a habit of giving, even with small amounts, helps build a long-term culture of altruism — and it would