Effective Altruism Forum
EA Forum

Hide table of contents
Comment Permalink
Vasco Grilo🔸2
0
0
Response to previous version

Thanks, Bradley, and welcome to the EA Forum! Strongly upvoted.

Given that it is unlikely that incorporating humidity would decrease heat-related mortality, my own view here is that this pushes current estimates towards a lower bound.

If adequately modelling humidity would increase heat deaths, I wonder whether it would also decrease cold deaths, such that the net effects is unclear.

In practice, these assumptions limit the ability to model things like extreme heat waves and heat domes, which can cause large fatality spikes (e.g. figure below from Washington State in 2021). Missing these features in some locations might be akin to missing almost all the possible heat related mortality in cooler climates.

As illustrated below, deaths from extreme cold and heat accounted for only a tiny fraction of the deaths from non-optimal temperature in 2015 in some countries, which attenuates the effect you are describing.

Reply
See in context
by Vasco Grilo🔸
25th Jul 20246 min read 9

25

Cause prioritizationForecastingCost-effectiveness analysisClimate changeClimate risk skepticismFermi estimate
Frontpage

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

Summary

  • I estimated the annual deaths and death rate from non-optimal temperature in 2030, 2050 and 2090 relative to the temperature in 1986-2005, by region and income group of The World Bank, for a global warming of 2.5 ºC by 2100. I concluded the annual deaths from non-optimal temperature from 1986-2005 to 2050:
    • Increase in low and lower middle income countries, and decrease in upper middle and high income countries.
    • Globally, increase by 138 k, which is:
      • 0.224 % of all the deaths in 2023.
      • 1.71 % of all the deaths from air pollution in 2021.
      • 3.05 times the annual deaths from natural disasters from 2010 to 2019.
  • It seems to me there is still significant uncertainty about whether global warming is good/bad from the point of view of changing global deaths from non-optimal temperature.
  • I suspect the indirect effects of global warming are very overrated.
  • I determine a cost-effectiveness for stratospheric aerosol injection (SAI) of 2.95*10^-4 DALY/$, which is:

Future deaths from non-optimal temperature

I estimated the annual deaths and death rate from non-optimal temperature in 2030, 2050 and 2090 relative to the temperature in 1986-2005 by region and income group of The World Bank. I relied on:

  • Data on the annual death rate from non-optimal temperature relative to the temperature in 1986-2005, accounting for changes in cold and heat-related mortality, for “2°C of warming by mid-century, and 2.5°C by 2100” from Human Climate Horizons. This was the 39th percentile global warming according to the Metaculus’ community on 15 July 2024, i.e. slightly better than the median scenario.
  • Data on medium population projections from the World Population Prospects.

The data and results are in this Sheet. Below I only show results for the annual deaths and death rate from non-optimal temperature in 2050 relative to the temperature in 1986-2005. I chose that year because 2030 felt too early to capture concerns about a warming world, and 2090 too late for predictions to be reliable.

For a global warming of 2 ºC by 2050, I concluded the annual deaths from non-optimal temperature:

  • Decrease in:
    • 4 regions, Europe and Central Asia, North America, East Asia and Pacific, and Latin America and Caribbean.
    • 2 income groups, high and upper middle income countries.
  • Increase in:
    • 3 regions, Middle East and North Africa, Sub-Saharan Africa, and South Asia.
    • 2 income groups, low and lower middle income countries.
  • Globally, increase by 138 k, which is:
    • 0.224 % (= 138*10^3/(61.65*10^6)) of all the deaths in 2023.
    • 1.71 % (= 138*10^3/(8.08*10^6)) of all the deaths from air pollution in 2021.

    • 10.9 % of the sum of the absolute values of all entities[1]. So it seems to me there is still significant uncertainty about whether global warming is good/bad from the point of view of changing global deaths from non-optimal temperature.

    • 3.05 (= 138*10^3/(45.3*10^3)) times the annual deaths from natural disasters from 2010 to 2019.
    • Possibly an underestimate of the expected deaths because the mean global warming by 2100 is arguably higher than 2.5 ºC. In any case, I suppose the conclusions suggested by the above qualitatively hold.

It is also unclear to me whether global warming has so far increased/decreased global deaths from non-optimal temperature. Zao et al. (2021) found a decrease studying 750 locations in 43 countries:

“From 2000–03 to 2016–19, the global cold-related excess death ratio changed by −0·51 percentage points (95% eCI −0·61 to −0·42) and the global heat-related excess death ratio increased by 0·21 percentage points (0·13–0·31), leading to a net reduction in the overall ratio.”

The disease burden per capita of non-optimal temperature has been decreasing.

On this overall topic, I liked Our World in Data’s (OWID's) 2 posts on how many people die from extreme temperatures, and how this could change in the future. Particularly, the graph below showing global warming mostly affects lower and lower middle income countries.

https://ourworldindata.org/grapher/change-heat-deaths-gdp

I shared a draft of this post with Johannes Ackva. He commented that the sophisticated case for caring about climate change has to do with the indirect effects of global warming, like increased risk of conflicts. I suspect such indirect effects are very overrated:

  • I got this sense from skimming John Halstead’s report Climate Change & Longtermism.
  • The direct effects have been very overrated. Some argued (and maybe still argue) more than 1.5 ºC of global warming by 2100 would be catastrophic, whereas I think it is more accurate to say it may well be positive/negative (in expectation).
  • I believe the sign uncertainty of the indirect effects, i.e. whether they are good/bad (in expectation), will tend to be larger than that of the direct effects, and this already seems quite large to me.
  • Climate change might have indirect cascade effects, but so do other problems, and it is unclear to me whether the magnitude of the direct effects as a fraction of that of the indirect effects is specially small for climate change.
  • Appealing to indirect cascade effects or other known unknowns feels a little like a regression to the inscrutable, which is characterised by the following pattern:
    • Arguments for high risk initially focus on aspects which are relatively better understood (e.g. changes in the deaths from non-optimal temperature).
    • Further analysis frequently shows the risk from such aspects has been overestimated, and is in fact quite low.
    • Then discussions move to more poorly understood aspects of the risk (e.g. global warming might increase the risk of conflicts).
  • The direct effects just look too small to cause significant indirect effects. For instance, the overall death rate in India is predicted to be 0.880 % in 2050, whereas the death rate from non-optimal temperature is only predicted to increase 0.006 pp from 1986-2005 to 2050, i.e. just 0.682 % (= 6*10^-5/0.00880) of the overall death rate projected for 2050.

Cost-effectiveness of stratospheric aerosol injection

SAI is “a proposed method of solar geoengineering (or solar radiation modification [SRM]) to reduce global warming. This would introduce aerosols into the stratosphere to create a cooling effect via global dimming and increased albedo, which occurs naturally from volcanic winter”.

I determined a cost-effectiveness for SAI of 2.95*10^-4 DALY/$ (= 6.44*10^6/(21.8*10^9)). I got this from the ratio between:

  • An annual benefit of SAI of 6.44 MDALY (= 138*10^3*46.7), which I obtained multiplying:
    • My estimate for the increase in the annual deaths from non-optimal temperature from 1986-2005 to 2050 for 2.5 ºC of global warming by 2100 of 138 k.
    • A disease burden per death from non-optimal temperature of 46.7 DALY (= 1.68*10^6/(36.0*10^3)), which is the ratio between the disease burden and deaths from environmental heat and cold exposure in 2021 from the Global Burden of Disease study (GBD) of 1.68 MDALY and 36.0 k.
  • An annual cost of SAI of 21.8 G$/year (= 1*21.8*10^9), which I computed from the product between:
    • 1 ºC (= 2.5 - 1) of global cooling, which I guess would eliminate the deaths mentioned just above by bringing global warming to 1.5 ºC by 2100.
    • The annual cost of SAI per degree of cooling from Smith 2020 of 18 G 2020-$/year/ºC, i.e. 21.8 G$/year/ºC (= 18*10^9*1.21).

My estimate for the cost-effectiveness of SAI is:

I speculate a leveraged intervention like research on or advocacy for SAI can be up to 100 times as cost-effective as directly implementing it, which would bring the cost-effectiveness up to 0.0295 DALY/$ (= 100*2.95*10^-4), i.e. roughly the same level as that of CCF. However, I do not see how such leveraged interventions could be more cost-effective than the best animal welfare interventions.

Acknowledgements

Thanks to Johannes Ackva for feedback on the draft. Johannes noted he is “not really the right person to review the climate science part of this” nor “a peer reviewer in the sense of having reviewed everything thoroughly”.

  1. ^

     This metric would be 1 if annual deaths increased in every single entity, in which case sign uncertainty would be minimal.

25

0
2

Reactions

0
2
Mentioned in

More posts like this

Comments8


Sorted by
New & upvoted
Click to highlight new comments since:
jackva12
2
0
Response to previous version

Thanks for this, Vasco, thought-provoking as always!

I do not have too much time to discuss this, but I want to point out that I am pretty unconvinced by the argument for why indirect effects should be easy to discount by the type of the argument you make. So, to be clear, I am not arguing that the conclusion is necessarily wrong (I think this would require stronger arguments), but rather that the argumentation strategy here does not work.

As far as I understand it your argument for "indirect effects cannot be very large" is something like:

1. Deaths from heat are very overstated and not that significant and they capture the most important direct effect.

2. A claim for strong indirect effects is then implausible because onto that direct effect you need to sequence a bunch of very uncertain additional effects with uncertain signs.

Insofar as is this a correct representation of your argument -- please let me know -- I think it has a couple of problems:

1.

a. Dying from heat stress is a very extreme outcome and people will act in response to climate change much earlier than dying. For example, before people die from heat stress, they might abandon their livelihoods and migrate, maybe in large numbers.

b. More abstractly, the fact that an extreme impact outcome (heat death) is relatively rare is not evidence for low impact in general. Climate change pressures are not like a disease that kills you within days of exposure and otherwise has no consequence.

2.

a. You seem to suggest we are very uncertain about many of the effect signs. I think the basic argument why people concerned about climate change would argue that changes will be negative and that there be compounding risks is because natural and human systems are adapted to specific climate conditions. That doesn't mean they cannot adapt at all, but that does mean that we should expect it is more likely that effects are negative, at least as short-term shocks, than positive for welfare. Insofar as you buy a story where people migrate because of climate impacts, for example, I don't think it is unreasonable to say that increased migration pressures are more likely to increase tensions than to reduce them, etc.

b. I think a lot of the other arguments on the side of "indirect risks are low" you cite are ultimately of the form (i) "indirect effects in other causes are also large" or (ii) "pointing to indirect effects make things inscrutable and unverifiable".  (i) might be true but doesn't matter, I think, for the question of whether warming is net-bad and (ii) is also true, but does nothing by itself on whether those indirect effects are real -- we can live in a world where indirect effects are rhetorically abused and still exist and indeed dominate in certain situations!
 

Reply
Vasco Grilo🔸2
0
0
Response to previous version

Thanks for this, Vasco, thought-provoking as always!

Likewise! Thanks for the thoughtful comment.

Insofar as is this a correct representation of your argument

It seems like a fair representation.

a. Dying from heat stress is a very extreme outcome and people will act in response to climate change much earlier than dying. For example, before people die from heat stress, they might abandon their livelihoods and migrate, maybe in large numbers.

b. More abstractly, the fact that an extreme impact outcome (heat death) is relatively rare is not evidence for low impact in general. Climate change pressures are not like a disease that kills you within days of exposure and otherwise has no consequence.

Agreed. However:

  • I think migration will tend to decrease deaths because people will only want to migrate if they think their lives will improve (relative to the counterfactual of not migrating).
  • The deaths from non-optimal temperature I mentioned are supposed to account for all causes of death, not just extreme heat and cold. According to GBD, in 2021, deaths from environmental heat and cold exposure were 36.0 k (I guess this is what you are referring to by heat stress), which was just 1.88 % (= 36.0*10^3/(1.91*10^6)) of the 1.91 M deaths from non-optimal temperature. My post is about how these 1.91 M deaths would change.

a. You seem to suggest we are very uncertain about many of the effect signs. I think the basic argument why people concerned about climate change would argue that changes will be negative and that there be compounding risks is because natural and human systems are adapted to specific climate conditions. That doesn't mean they cannot adapt at all, but that does mean that we should expect it is more likely that effects are negative, at least as short-term shocks, than positive for welfare.

This makes sense. On the other hand, one could counter global warming will be good because:

  • There are more deaths from low temperature than from high temperature.
  • The disease burden per capita from non-optimal temperature has so far been decreasing (see 2nd to last graph).

b. I think a lot of the other arguments on the side of "indirect risks are low" you cite are ultimately of the form (i) "indirect effects in other causes are also large" or (ii) "pointing to indirect effects make things inscrutable and unverifiable".  (i) might be true but doesn't matter, I think, for the question of whether warming is net-bad and (ii) is also true, but does nothing by itself on whether those indirect effects are real -- we can live in a world where indirect effects are rhetorically abused and still exist and indeed dominate in certain situations!

Agreed. I would just note that i) can affect prioritisation across causes.

Reply
jackva4
0
0
Response to previous version

Thanks!

I think I remain confused as to what you mean with "all deaths from non-optimal temperature".

It is clear that the data source you cite (GBD, focused on current deaths) will not feature nor proxy what people concerned about compounding risks from climate are concerned about.

So to me it seems you are saying "I don't trust arguments about compounding risks and the data is evidence for that" whereas the data is inherently not set up to include that concern and does not really speak to the arguments that people most concerned about climate risk would make.

As said before, I think it is fine to say "I don't trust arguments about compounding risks" and I am probably with you there to a large degree at least compared to people most concerned about this, but I don't think the data from GBD is additional evidence for that mistrust, as far as I can tell.

By crude analogy, if you believed that COVID restrictions had a big toll on the young and this will affect long-run impacts somehow, pointing to few COVID deaths amongs this age cohort would not be evidence against this concern.

Reply
Vasco Grilo🔸4
0
0
Response to previous version

I think I remain confused as to what you mean with "all deaths from non-optimal temperature".

I mean the difference between the deaths for the predicted and ideal temperature. From OWID:

Schematic figure showing how temperature relates to mortality risk. Risk increases at extreme cold and hot temperatures.

The deaths from non-optimal temperature are supposed to cover all causes (temperature is a risk factor for death rather than a cause of death in GBD), not just extreme heat and cold (which only account for a tiny fraction of the deaths; see my last comment). I say "supposed" because it is possible the mortality curves above are not being modelled correctly, and this applies even more to the mortality curves in the future.

So to me it seems you are saying "I don't trust arguments about compounding risks and the data is evidence for that" whereas the data is inherently not set up to include that concern and does not really speak to the arguments that people most concerned about climate risk would make.

My understanding is that (past/present/future) deaths from non-optimal temperature are supposed to include conflict deaths linked to non-optimal temperature. However, I am not confident these are being modelled correctly.

I was not clear, but in my last comment I mostly wanted to say that deaths from non-optimal temperature account for the impact of global warming not only on deaths from extreme heat and cold, but also on cardiovascular or kidney disease, respiratory infections, diabetes and all others (including conflicts). Most causes of death are less heavy-tailed than conflict deaths, so I assume we have a better understanding of how they change with temperature.

Reply
jackva4
0
0
Response to previous version

Ok, so I think we converge pretty much then -- essentially what I am saying is that people concerned about compounding risks would argue that these are not modeled correctly in GBD and that there is much more uncertainty there (and that the estimate is probably an underestimate, from the perspective of taking the compounding risk view seriously).

Reply
Vasco Grilo🔸2
0
0
Response to previous version

Makes sense. Just to clarify, the data on deaths and disease burden from non-optimal temperature until now are from GBD, but the projections for the future death rates from non-optimal temperature are from Human Climate Horizons.

Reply
Bradley Wilson10
0
0
Response to previous version

I'm not sure it would change your underlying conclusions, but for what it's worth I took a quick look at the data/methods from Human Climate Horizons and think that it is likely an underestimate of future heat-related mortality. 

This paper underlies the Human Climate Horizons/OWID estimates. A few quick assumptions I think are worth highlighting:

1. The authors (by admission) do not consider the effects of humidity. Accurate humidity data is much harder to come by than temperature, and it's often not included in the down-scaled versions of the climate models, though some exceptions apply. However, it's potentially quite relevant for heat-related mortality, though there is some debate in the literature. Given that it is unlikely that incorporating humidity would decrease heat-related mortality, my own view here is that this pushes current estimates towards a lower bound. 

2. The authors use climate model data that are downscaled with a technique called Bias-Correction Spatial Disaggregation (BCSD). This approach has two key assumptions in it: (1) it assumes that the relative spatial patterns in the training data will remain constant under future climate change, and (2) it is calibrated to monthly data, not day to day changes, which has the effect of dampening extreme values. In practice, these assumptions limit the ability to model things like extreme heat waves and heat domes, which can cause large fatality spikes (e.g. figure below from Washington State in 2021). Missing these features in some locations might be akin to missing almost all the possible heat related mortality in cooler climates. 

Again, I don't think these are significant enough to fundamentally change your conclusions, but I do think it's worth highlighting that these types of results can be quite sensitive to specifics of the climate modeling approaches that are used. 

Reply
Vasco Grilo🔸2
0
0
Response to previous version

Thanks, Bradley, and welcome to the EA Forum! Strongly upvoted.

Given that it is unlikely that incorporating humidity would decrease heat-related mortality, my own view here is that this pushes current estimates towards a lower bound.

If adequately modelling humidity would increase heat deaths, I wonder whether it would also decrease cold deaths, such that the net effects is unclear.

In practice, these assumptions limit the ability to model things like extreme heat waves and heat domes, which can cause large fatality spikes (e.g. figure below from Washington State in 2021). Missing these features in some locations might be akin to missing almost all the possible heat related mortality in cooler climates.

As illustrated below, deaths from extreme cold and heat accounted for only a tiny fraction of the deaths from non-optimal temperature in 2015 in some countries, which attenuates the effect you are describing.

Reply
More from Vasco Grilo🔸
46
Is it 3 Years, or 3 Decades Away? Disagreements on AGI Timelines
Vasco Grilo🔸
· · 2m read
1
1
93
Yes, Shrimp Matter
Vasco Grilo🔸
· · 8m read
10
10
41
Wild animal suffering is super neglected compared with factory-farming, and this is super neglected compared with conflicts, epidemics/pandemics, and global warming
Vasco Grilo🔸
· · 7m read
4
4
View more
Curated and popular this week
 ·  · 5m read
 · 
This work has come out of my Undergraduate dissertation. I haven't shared or discussed these results much before putting this up.  Message me if you'd like the code :) Edit: 16th April. After helpful comments, especially from Geoffrey, I now believe this method only identifies shifts in the happiness scale (not stretches). Have edited to make this clearer. TLDR * Life satisfaction (LS) appears flat over time, despite massive economic growth — the “Easterlin Paradox.” * Some argue that happiness is rising, but we’re reporting it more conservatively — a phenomenon called rescaling. * I test rescaling using long-run German panel data, looking at whether the association between reported happiness and three “get-me-out-of-here” actions (divorce, job resignation, and hospitalisation) changes over time. * If people are getting happier (and rescaling is occuring) the probability of these actions should become less linked to reported LS — but they don’t. * I find little evidence of rescaling. We should probably take self-reported happiness scores at face value. 1. Background: The Happiness Paradox Humans today live longer, richer, and healthier lives in history — yet we seem no seem for it. Self-reported life satisfaction (LS), usually measured on a 0–10 scale, has remained remarkably flatover the last few decades, even in countries like Germany, the UK, China, and India that have experienced huge GDP growth. As Michael Plant has written, the empirical evidence for this is fairly strong. This is the Easterlin Paradox. It is a paradox, because at a point in time, income is strongly linked to happiness, as I've written on the forum before. This should feel uncomfortable for anyone who believes that economic progress should make lives better — including (me) and others in the EA/Progress Studies worlds. Assuming agree on the empirical facts (i.e., self-reported happiness isn't increasing), there are a few potential explanations: * Hedonic adaptation: as life gets
 ·  · 38m read
 · 
In recent months, the CEOs of leading AI companies have grown increasingly confident about rapid progress: * OpenAI's Sam Altman: Shifted from saying in November "the rate of progress continues" to declaring in January "we are now confident we know how to build AGI" * Anthropic's Dario Amodei: Stated in January "I'm more confident than I've ever been that we're close to powerful capabilities... in the next 2-3 years" * Google DeepMind's Demis Hassabis: Changed from "as soon as 10 years" in autumn to "probably three to five years away" by January. What explains the shift? Is it just hype? Or could we really have Artificial General Intelligence (AGI)[1] by 2028? In this article, I look at what's driven recent progress, estimate how far those drivers can continue, and explain why they're likely to continue for at least four more years. In particular, while in 2024 progress in LLM chatbots seemed to slow, a new approach started to work: teaching the models to reason using reinforcement learning. In just a year, this let them surpass human PhDs at answering difficult scientific reasoning questions, and achieve expert-level performance on one-hour coding tasks. We don't know how capable AGI will become, but extrapolating the recent rate of progress suggests that, by 2028, we could reach AI models with beyond-human reasoning abilities, expert-level knowledge in every domain, and that can autonomously complete multi-week projects, and progress would likely continue from there.  On this set of software engineering & computer use tasks, in 2020 AI was only able to do tasks that would typically take a human expert a couple of seconds. By 2024, that had risen to almost an hour. If the trend continues, by 2028 it'll reach several weeks.  No longer mere chatbots, these 'agent' models might soon satisfy many people's definitions of AGI — roughly, AI systems that match human performance at most knowledge work (see definition in footnote). This means that, while the compa
 ·  · 4m read
 · 
SUMMARY:  ALLFED is launching an emergency appeal on the EA Forum due to a serious funding shortfall. Without new support, ALLFED will be forced to cut half our budget in the coming months, drastically reducing our capacity to help build global food system resilience for catastrophic scenarios like nuclear winter, a severe pandemic, or infrastructure breakdown. ALLFED is seeking $800,000 over the course of 2025 to sustain its team, continue policy-relevant research, and move forward with pilot projects that could save lives in a catastrophe. As funding priorities shift toward AI safety, we believe resilient food solutions remain a highly cost-effective way to protect the future. If you’re able to support or share this appeal, please visit allfed.info/donate. Donate to ALLFED FULL ARTICLE: I (David Denkenberger) am writing alongside two of my team-mates, as ALLFED’s co-founder, to ask for your support. This is the first time in Alliance to Feed the Earth in Disaster’s (ALLFED’s) 8 year existence that we have reached out on the EA Forum with a direct funding appeal outside of Marginal Funding Week/our annual updates. I am doing so because ALLFED’s funding situation is serious, and because so much of ALLFED’s progress to date has been made possible through the support, feedback, and collaboration of the EA community.  Read our funding appeal At ALLFED, we are deeply grateful to all our supporters, including the Survival and Flourishing Fund, which has provided the majority of our funding for years. At the end of 2024, we learned we would be receiving far less support than expected due to a shift in SFF’s strategic priorities toward AI safety. Without additional funding, ALLFED will need to shrink. I believe the marginal cost effectiveness for improving the future and saving lives of resilience is competitive with AI Safety, even if timelines are short, because of potential AI-induced catastrophes. That is why we are asking people to donate to this emergency appeal
Relevant opportunities
5
Q2 AI Forecasting Benchmark: $30,000 in Prizes
christian
· · 1m read
0
0
15
Open call: "Existential risk of AI: technical conditions"
miller-max
· · 1m read
0
0
113
AI Moral Alignment: The Most Important Goal of Our Generation
Ronen Bar
· · 10m read
31
31
View more