Such a great question! If we were to do a more rigorous cost-effectiveness evaluation, this would be one of the first things we'd try to add to the model.
extreme weather and accidents (e.g. falling out of trees)
There's even less research on the relative amounts of suffering from different causes of death, but rabies is a pretty good candidate for one of the worst deaths out there, because it causes extreme suffering over several days. (It's tragically convenient that humans die of rabies and exhibit similar symptoms, so we can back up our guesses with first-hand reports and observations of behavior in a species we know well.) Canine distemper might be worse, because it also causes intense suffering but can last for weeks.
But most of the likely most common causes of raccoon deaths act much more quickly (some almost instantaneously), it seems like a pretty safe bet that rabies is substantially worse than most other raccoon deaths.
Executive summary: This exploratory post argues that wild animal welfare science could yield highly cost-effective interventions to reduce suffering, using a rough cost-effectiveness analysis of vaccinating raccoons against rabies in Texas to show that such programs are already promising and likely represent the low end of what future interventions could achieve.
Key points:
Wild animal welfare science holds promise for identifying scalable, cost-effective ways to reduce suffering in nature, though the field is still speculative and underdeveloped.
Rabies vaccination programs for wild mammals, like raccoons in Texas, offer a real-world example of an intervention that benefits animals, is large-scale, and relatively affordable.
A back-of-the-envelope estimate suggests the program may prevent a raccoon death from rabies for around $15–20, making it surprisingly cost-effective even though it wasn’t designed with animal welfare as the primary goal.
The analysis relies on many assumptions and uncertain data, and the authors caution that results could be off by orders of magnitude — but the point is to demonstrate plausibility, not precision.
As the field matures, the authors believe future interventions designed specifically for wild animal welfare could be dramatically more cost-effective, through better targeting, novel technologies, and economies of scale.
They advocate for increased investment in wild animal welfare science, arguing that foundational research and institutional support are needed to unlock the full potential of this space.
This comment was auto-generated by the EA Forum Team. Feel free to point out issues with this summary by replying to the comment, and contact us if you have feedback.
This post explains why we think wild animal welfare science could lead to the development of large-scale, highly cost-effective interventions to reduce wild animal suffering. To check our rationale against a historical, real-world example with data, we do a back-of-the-envelope estimation of the cost-effectiveness of vaccinating raccoons in Texas against rabies.
[...]
Given that the program costs about $2.3 million dollars per year, that would mean the average cost of preventing one raccoon from dying of rabies (and suffering greatly along the way) would be $15.64 [0.0639 deaths from rabies averted per $ (= 1/15.64)].
I like that you estimated the cost-effectiveness, but I do not think it illustrates "animal welfare science could lead to the development of large-scale, highly cost-effective interventions". The Shrimp Welfare Project's (SWP's) Humane Slaughter Initiative (HSI) averts the suffering involved in the air asphyxiation of 15 k shrimp per $, thus helping 235 k (= 15*10^3*15.64) times as many animals per $ as vaccinating raccoons. From Table 8.6 of Bob Fischer's book about comparing welfare across species, the welfare range of shrimp is 18.2 % (= 0.080/0.44) of that of pigs. Assuming the welfare range of raccoons equals that of pigs, which I guess overestimates the welfare range of raccoons, the welfare range of shrimp is 18.2 % that of raccoons. In this case, HSI would help 42.8 k (= 235*10^3*0.182) as many animals weighted by welfare range per $ as vaccinating raccoons.
There are also other interventions which look way more cost-effective than vaccinating raccoons accounting for effects on soil animals, even relying on my preferred welfare ranges proportional to the square root of the number of neurons. I estimate the Centre for Exploratory Altruism Research’s (CEARCH’s) High Impact Philanthropy Fund (HIPF) has a cost-effectiveness accounting for target beneficiaries, and soil ants, termites, springtails, mites, and nematodes of 70.6 kQALY/$ due to decreasing the living time of those animals by 5.07 billion animal-years per $. Even if preventing one raccoon from dying of rabies was as good as 10 QALY (10 fully happy human-years), the cost-effectiveness of vaccinating raccoons accounting only for target beneficiaries would be 0.639 QALY/$ (= 10/15.64), 9.05*10^-6 (= 0.639/(70.6*10^3)) of my estimate for the cost-effectiveness of funding HIPF accounting for target beneficiaries and soil animals.
@Cameron Meyer Shorb 🔸, has Wild Animal Initiative considered doing or funding research on soil animals?
I like that you estimated the cost-effectiveness, but I do not think it illustrates "animal welfare science could lead to the development of large-scale, highly cost-effective interventions"
I was worried I didn't articulate this claim clearly enough in the original post, so I appreciate you giving me the chance to clarify!
I did not mean to say "This is highly cost-effective relative to other animal welfare interventions" or "This is about how cost-effective I expect wild animal welfare interventions to be."
I was aiming for something more like: "I don't think it's crazy to think we could develop highly cost-effective wild animal welfare interventions, because here's an example of a program that wasn't even optimized for wild animal welfare but still seemed to be doing likely-good things fairly affordably."
Wild Animal Initiative's whole deal is advancing research that we think will lead to interventions that are orders of magnitude more scalable and more cost-effective than what's possible now. Some people seem to think that's hopeless, in part because it seems so far beyond what we can imagine now. We disagree. The goal of this post wasn't to provide definitive proof for every part of that argument, but rather to provide an empirical reality check. Given how early we are in the history of wild animal welfare science, I think it's a pretty encouraging sign that we're already doing plausibly-helpful things that are within a few orders of magnitude of competitive cost-effectiveness.
[P.S. Still planning to reply to your point about soil animals, but ran out of time today.]
I was aiming for something more like: "I don't think it's crazy to think we could develop highly cost-effective wild animal welfare interventions, because here's an example of a program that wasn't even optimized for wild animal welfare but still seemed to be doing likely-good things fairly affordably."
I understand that. However, advocating for interventions improving the lives of wild animals seems way less promising that for popular interventions which cost-effectively change land use, such as cost-effectively saving human lives (to decrease the number of soil animals if these have negative lives), or increasing forest area (to increase the number of soil animals if these have positive lives):
Funding HIPF is also not optimised for increasing the welfare of wild animals, and I estimate it increases the welfare of soil animals 110 k (= 1/(9.05*10^-6)) times as cost-effectively as vaccinating raccoons increases the welfare of these even if preventing one raccoon from dying of rabies was as good as 10 QALY.
The cost-effectiveness of advocating for an intervention is the cost-effectiveness of funding the intervention times the money moved to the intervention as a fraction of the spending advocating for it. Fundraising for saving human lives, and increasing forest area is much easier than for improving the conditions of soil animals. So advocating for the former will be more cost-effective than for the latter even if the cost-effectiveness of funding them directly was the same.
[P.S. Still planning to reply to your point about soil animals, but ran out of time today.]
Are you still planning to reply to my points about soil animals?
This post presents the executive summary from Giving What We Can’s impact evaluation for 2025. At the end of this post we share links to more information, including the full report and...
I used AI to fix transcription errors, rerrarange the ideas, and suggest tweaks to the title and some sentences.
Three of the most exciting projects to come out of EA in recent years are, in a vague sense, CEA spinouts:
* Kairos is directly a spinout of CEA and now handles most support for university AI safety groups. Basically everyone I've found who knows them is really excited about what they do
* NEST is an opinionated ideas-fi...
This post explains why we think wild animal welfare science could lead to the development of large-scale, highly cost-effective interventions to reduce wild animal suffering. To check our rationale against a historical, real-world example with data, we do a back-of-the-envelope estimation of the cost-effectiveness of vaccinating raccoons in Texas against rabies.
What makes early research in wild animal welfare science so exciting — or frustrating, depending on who you ask — is that we don't know exactly where it will lead. We’re asking questions precisely because we don’t know the answers.
Still, it’s reasonable to ask whether this will all be worth it. Wild systems are complicated and, by definition, not entirely under human control. Many present-day efforts to help wild animals — treating severe injuries, for instance — are highly resource intensive and hard to scale. Does that mean wildlife management aimed at improving welfare is hopelessly expensive?
We don’t think so, for a few reasons: There are lots of wild animals; humans are already affecting them so much that minor policy changes could have huge effects; the historic neglectedness of this issue means that there are probably some relatively easy wins available; scaling current technologies can bring down costs; and inventing new technologies can open new possibilities.
We think these reasons are compelling, but we also know that they’re speculative. To check our rationale against some real-world data, we thought we’d quickly run the numbers on the cost-effectiveness of a wildlife management program that could work as a proxy for the kinds of interventions we’d like to see more of.
Rabies vaccines for wild mammals
Rabies is a viral disease that is nearly 100% fatal once symptoms appear. For humans and many other mammals it affects, it seems to be among the more painful causes of death, causing progressively intense symptoms over several days, including hyperactivity, violent convulsions, inability to swallow, and partial paralysis.
With the widespread vaccination of people and pets, rabies now resides primarily in wild animal populations. To reduce its prevalence, the U.S. Department of Agriculture partners with health departments in several states to distribute oral rabies vaccines (ORV) that are safe and effective for various terrestrial mammal species. Millions of doses are dropped from helicopters into forests and other natural areas each year. Although the program seems primarily motivated by human interests, it also directly benefits the wild animals being vaccinated.
We chose to highlight this example because it has many of the hallmarks of a promising intervention: It targets a source of intense suffering, it can potentially help a large number of animals (i.e., its benefits aren’t limited to threatened or endangered species), it can be deployed over large areas with relatively little human labor, and it uses technology that’s already affordable enough to be government-funded.
So when we refer to large-scale, cost-effective interventions, this vaccination program is the kind of thing we’re thinking of. But we also think it’s on the lower end of the cost-effectiveness spectrum for what will ultimately be possible. As wild animal welfare science advances, we expect to find ways to improve cost-effectiveness by several orders of magnitude, including by prioritizing the most tractable issues, designing interventions with wild animal welfare specifically in mind, and refining and scaling technologies and methodologies.
A back-of-the-envelope estimate of cost-effectiveness
We treated this as a rough thought experiment, not a rigorous research project. We narrowed our scope to raccoons in Texas, and, in order to piece together the little data we could find, we had to make a lot of assumptions. You can see all our calculations, assumptions, and references in this Guesstimate model, but here’s an overview of the key points:
We estimate that about 35% of raccoon deaths in Texas were attributable to rabies before the ORV program started in 1995. If Texas had a stable population of 3 million raccoons with an average lifespan of 2.5 years (i.e., 40% of them die each year), then about 420,000 raccoons were dying of rabies in Texas each year.
We couldn’t find any data on outcomes specific to Texas, so let’s assume the program reduced rabies by 70% by 2021, which is how much rabies declined nationwide during the same period (a conservative estimate, because it includes some areas with ORV programs and many without). That would mean that, by the end of the period between 1995–2021, 294,000 fewer raccoons were dying from rabies in Texas each year. Because this decline was roughly linear, we’ll say that the average number of rabies deaths averted per year (RDA/yr) over the period was simply the average of the start point (0 RDA/yr) and the end point (294,000 RDA/yr): 147,000 RDA/yr.
Given that the program costs about $2.3 million dollars per year, that would mean the average cost of preventing one raccoon from dying of rabies (and suffering greatly along the way) would be $15.64. Accounting for the asymmetry in our uncertainties bumps the price tag up to around $20, because there are more ways for this benefit to cost more than there are for it to cost less (see our model, which produces slightly different results each time because it introduces random variation into the calculation).
Figure 1. Summary of cost-effectiveness calculations. See this Guesstimate model for references and more detail.
Just a starting point
It’s important to stress just how rough that estimate is. Our model is intentionally crude, and the data we put into it left much to be desired. Our estimate could easily be off by a couple orders of magnitude in either direction.
As statisticians love to say: “All models are wrong, but some are useful.” The purpose of this model isn’t to determine the exact cost of sparing a raccoon from rabies, or to compare ORV against other wild animal welfare interventions — it’s to check whether it’s reasonably possible that wild animal welfare science could lead to the development of highly cost-effective interventions. Whatever the true value of ORV programs is today, we should expect it to be on the low end of what could be accomplished in the future. Research can identify more promising targets, find points of higher leverage, improve the effectiveness of technology, and reveal ways to lower the costs of production. If nothing else, interventions are likely to be much more effective at helping wild animals if they are designed specifically for that purpose, as opposed to incidentally benefiting animals insofar as it advances human health and economic priorities.
This is why we think that supporting the growing field of wild animal welfare science is a critical step to unlocking more ways to help wild animals. Actually doing the necessary research — and investing in the people and institutions who can sustain knowledge production — will not be as immediately rewarding as dumping vaccine pellets out of helicopters. But we think it will pay off in the long run, given how effective existing programs might be and how much room they have to improve.
29
More posts like this
75
Cost-Effectiveness of THL’s Corporate Cage-free Campaigns
Looks neat! How much worse do you think a rabies death would be compared to the other death the raccoon would have experienced?
Such a great question! If we were to do a more rigorous cost-effectiveness evaluation, this would be one of the first things we'd try to add to the model.
As is too often the case with cause-specific mortality, we couldn't find any great data on this. But based on our quick read of the literature and our general knowledge of natural history, @Simon Eckerström Liedholm and I think the leading candidates for counterfactual causes of raccoon death might be:
There's even less research on the relative amounts of suffering from different causes of death, but rabies is a pretty good candidate for one of the worst deaths out there, because it causes extreme suffering over several days. (It's tragically convenient that humans die of rabies and exhibit similar symptoms, so we can back up our guesses with first-hand reports and observations of behavior in a species we know well.) Canine distemper might be worse, because it also causes intense suffering but can last for weeks.
But most of the likely most common causes of raccoon deaths act much more quickly (some almost instantaneously), it seems like a pretty safe bet that rabies is substantially worse than most other raccoon deaths.