A few weeks ago, an open letter appeared calling on Effective Altruists to go beyond the cage-free campaigns and shift the focus from "welfare reforms" to promoting animal-free diets, especially in the wake of potentially very large amounts of funding flowing towards animal causes.
The original letter is here: https://beyondcagefree.org/ and an EA forum blog post by one of the organisers (@Anita Krajnc) can be found here. I also listened to Anita's interview in A Fork in the Road, by Vegan Report.
This is my humble attempt to write a reply, representing only my views as an effective altruist, but not others' and certainly not the EA community. The arguments of the letter, if I understand them correctly, are:
- Claim 1: Cage-free or welfare reforms have little or potentially even negative impact on other causes, including climate change and zoonotic diseases.
- Claim 2: As asserted in PETA's white paper on cage-free reforms, cage-free reforms make animal conditions "either marginally improved, unchanged, or, in some respects, even worsened".
- Claim 3: Total amount of animal suffering should not be the only variable, and we should also account for other moral theories (e.g. deontology or virtue ethics) as well as other metrics beyond suffering, such as animal rights.
This leads to the thesis that devoting most resources to cage-free campaigns is misguided, and we should shift some or most of the efforts to promoting plant-based diets. The three recommendations they give are:
- Broaden the Effective Altruism philosophy beyond utilitarianism and adopt a new impact metric focused on reducing the total number of animals bred into the food system.
- Develop an “Action Plan for a Plant-Based Future” that addresses both supply and demand (citing Meatless Mondays).
- Develop a High Ambition Coalition for Plant-Based Food Systems Transformation.
Let me try to address the claims, as I see them. Perhaps the weakest claim is the second: I believe PETA's white paper is a strawman of the evidence. This is Claude Opus 5.5 summary of the scientific evidence:
The letter cites PETA’s white paper to claim cage-free gains are marginal at best. The evidence does not support this.
- The EU’s scientific advisory body reviewed the evidence in 2023. It found cages severely restrict movement and natural behaviours, and recommended ending them to improve hen welfare (EFSA 2023).
- The most detailed quantification finds each aviary hen avoids at least 275 hours of disabling pain and 2,300 hours of hurtful pain versus a battery cage (Welfare Footprint).
- Higher cage-free mortality was a transition effect. In mature systems, it no longer differs (Schuck-Paim et al. 2021).
Cage-free hens do suffer more keel fractures, pecking injuries and dust exposure (CSES 2015). But the letter itself concedes cage-free lets hens nest, forage, turn around and spread their wings. Deprivation of these behaviors is the largest source of pain in cages. A list of harms is not a weighing of harms. When the harms are weighed, cage-free is clearly better.
Other AI systems seem to arrive at similar conclusions. I am not an expert myself, so it is hard to really gauge the evidence. Still, to the best of my ability, cage-free campaigns do significantly reduce suffering, all else being equal.
On the other hand, it seems that, indeed, cage-free is marginally worse for the environment, mostly because animals move more and thus spend more energy on things other than producing meat/eggs. However, Effective Altruists try to account for the scale of the problem. This means comparing the scale of animal suffering improvement against the scale of climate change worsening.
Since we already covered the animal welfare improvement side, let's now compare the climate change worsening: global equivalent CO2 emissions were 57.7 Gt COâ‚‚e in 2024. Livestock emitted 6.2 Gt COâ‚‚e in 2015, about 12% of the 50 to 52 Gt global total (FAO 2023; summary). Chickens account for 9% of livestock emissions. By product, meat is two-thirds, milk 30%, and eggs the rest (same FAO report). So eggs emit roughly 0.21 Gt COâ‚‚e (~= 6.2 * (1-2/3-0.3)), about 0.5% of global emissions. Chickens overall are0.56 Gt COâ‚‚e.
Eggs cause about 0.2 Gt COâ‚‚e per year, roughly 0.4% of global emissions. Estimates of the extra emissions from going cage-free range from none (Pelletier 2017) to about 20% per kg (Leinonen et al. 2012; Dekker et al. 2011). Fully organic production would double emissions (Martinic et al. 2026). However, cage-free campaigns target barn and aviary systems, not organic. Even a full global switch would add under 0.1% to global emissions. Part of global production is already cage-free, so a realistic figure is closer to 0.05%.
The other large group of caged animals is breeding sows, kept in gestation and farrowing crates. Pigs cause 14% of livestock emissions, about 0.9 Gt CO₂e or 1.7% of the global total (FAO 2023). Crate bans only change how sows are housed. The fattening herd causes most pig emissions and is unaffected (Thoma et al. 2018). One life-cycle study modelled US farms moving sows from gestation stalls to group pens (Bandekar 2015). Sows in pens needed slightly less feed. Pens also needed about 65% more barn space, which partly offset that gain. The net result was about 1% lower emissions per kg of pig. However, only 58–60% of the model’s uncertainty simulations showed a reduction, so one could argue that the result is indistinguishable from 0. On the other hand, we found no comparable study for farrowing crates. Free-farrowing pens need more space and can raise piglet mortality, so a small increase is plausible. If crate-free housing changes pig emissions by 0–5%, the global effect is at most 0.08%. Together, moving all caged animals to cage-free would likely add under 0.2% to global emissions.
Climate impact is not just COâ‚‚ emissions. However, COâ‚‚e emissions might easily be the most important environmental impact. Additionally, historically the EA community has placed greater emphasis on animal welfare than on climate change EA Survey 2024, and for what it is worth, I agree with that prioritisation. I think it is a good testament to the EA community to have almost always placed animal welfare among the top 3 causes even when climate change has been more popular among the general public.Â
Perhaps the largest uncertainty is the impact on wild animal welfare. One could consider second-order effects on soil nematode wildlife, which, taken seriously, would violently swing the estimates of whether cage-free is indeed good or bad; see @Vasco Grillo's posts. However, it contains a lot of uncertainty and has a stronger dependence on fully embracing utilitarianism, which gives pause to many, including myself. The impact on vertebrate wildlife has also been discussed in this forum but seems hard to settle. My key takeaway is that, if we do not account for potential displacement of wild animals whose life net-positivity might be hard to ascertain, the climate impact of cage-free transitions is tiny on the large scale of things (which is good!).
On disease and use of antibiotics, the evidence of cages is mixed. A common assumption is that cage-free hens live outdoors, but most campaigns target barns and aviaries, where hens stay indoors. The main increase in pandemic risk comes from contact with wild birds. In the Netherlands, free-range farms were 6.3 times more likely to catch low-pathogenic bird flu than indoor farms (Bouwstra et al. 2017, cited in Gonzales et al. 2021). Indoor cage-free systems have their own trade-offs. Hens living on litter are in closer contact with their droppings. They get more parasites and bacterial infections, and more of them die (Lay et al. 2011). A survey of 405 German farms found that caged flocks needed fewer veterinary treatments, including antibiotics (Kreienbrock et al. 2007). But cages have problems too. In an EU survey of 3,768 flocks, caged hens were about five times more likely to carry Salmonella (Van Hoorebeke et al. 2010). Antibiotic resistance showed no clear link to housing type (Van Hoorebeke et al. 2011).
Cost is the final side effect highlighted in the letter that I think is worth discussing here. Cage-free eggs are more expensive to produce. Barn eggs cost about 17% more in the EU (Majewski et al. 2024). A US trial found 30–40% more (UC Davis). However, this cuts both ways again: it will make food less affordable in poor countries, but higher prices should also mechanically reduce the amount of eggs consumed, although not by much: a 10% price rise cuts consumption by roughly 3% (Andreyeva et al. 2010). If the full 17% cost increase reached shoppers, egg consumption would fall by about 5%.
Further, I think an often-overlooked impact of cage-free campaigns is that they might be shifting public views on animal welfare. The "End the cage age" campaign in Europe and Proposition 12 in California, for example, have created some public recognition that animals matter. By getting people to sign the petition, opinions might be changing and shifting in the direction that animals are sentient beings. This strategy could eventually create the same public consensus that currently exists on climate change, and lead to legal bans (Shapiro and Siu). This consensus seems rather necessary if we want to end factory farming, and signing a petition often feels a lot easier than going vegan. In other words: politically oriented cage-free campaigns might be inadvertently leveraging the foot-in-the-door strategy often advocated for by Tobias Lenaert; see How to create a vegan world.
Now, with respect to the focus on utilitarianism, I think the claim is a bit too strong. Effective altruists indeed tend to place significant emphasis on utilitarianism, among other things, because historically it has been more prescient in moral matters than alternative theories. However, Effective Altruism has also championed the idea of moral uncertainty, and Rethink Priorities has a moral parliament tool that allows many in the community to account for more than basic utilitarianism. This, unfortunately, does not necessarily give clear, settled answers. Perhaps the claim by the Beyond Cage Free campaign supporters is that it's not just animal suffering that matters, but also animal rights, and that animals do not deserve to be killed. For what it's worth, I think that is true, and many Effective Altruists would agree. I would love to see the end of factory farming, and let animals live nice lives.
However, the biggest disagreement is not about utilitarianism, but about tractability: if there were clear ways to effectively promote plant-based diets, the Effective Altruism community would likely be highly supportive of it. For example, there is an ongoing debate in the community about the extent to which alternative proteins could be such a candidate (here, here). So tractability is, in my humble opinion, the actual crux; if animal liberation advocates proved there were tractable ways of promoting plant-based diets, it would change many opinions. There is some pessimism that this is even possible, and the reason why cage-free campaigns have been a well-funded intervention: because at least here, we can get partial wins (all things accounted for). However, I (and I guess many in the EA community) would be excited to learn effective ways to promote plant-based diets, so please let us know if you find any!
Disclaimer: please note this only represents my views, not those of the EA community or my employer. I'd like to make this point super clear.
Hello.
Thanks for covering this. My take that corporate campaigns for cage-free layers could be good or bad for animals does not depend on caring about soil nematodes, or fully embracing utilitarianism. I think their effects on soil ants and termites could be much larger than those on chickens. So my take would not change if I attributed a moral weight of exactly 0 to soil springtails, mites, and nematodes. One can also be concerned about such campaigns decreasing the welfare of soil ants and termites a lot without fully embracing utilitarianism.
Here are my estimates for the cost-effectiveness of funding The Navigation Fund's cage-free campaigns fund assuming moral weight is proportional to "individual number of neurons"^"exponent".
As illustrated below, an exponent of 0.188 explains the sentience-adjusted welfare ranges in Bob Fischer's book very well. In agreement with the bottom left graph above, such exponent implies a 50 % chance of the campaigns increasing welfare. This results from very uncertain effects on soil animals being much larger than the effects on chickens, as it is illustrated in the bottom right graph above.
I agree that the strongest argument against cage-free campaigns is uncertain wild animal effects, and that we don't need to assign moral weight to nematodes for that.
That said, I'd be interested in your perspective on the utilitarianism consideration, because that seems more load-bearing both for PAMC and for myself. I feel personally responsible for the suffering of hens, both because I used to contribute to it and because factory farming wouldn't be happening without humanity. I care about tiny wild animals too, but I feel less of a duty towards them. Alongside Brian Tomasik's blog, your arguments have contributed to me shifting most of my animal welfare giving to invertebrates. However, for deontological reasons, I'm still glad when people donate to chicken welfare, even when I agree the overall utilitarian sign is uncertain. If you have a moral parliament or otherwise give deontology and/or special duties some weight, that's a reason to support corporate campaigns. What do you think?
Hello Ariel.
The effects on land use caused by changes in food consumption would also not happen without humanity. Do you think humanity is more responsible for causing cage-free layers to exist than the changes in land use caused by them replacing caged layers? If not, why would the effects on chickens be special relative to those on soil invertebrates resulting from the changes in land use?Â
Nice to be mentioned alongside Brian Tomasik. Have you considered supporting the interventions which increase/decrease agricultural land the most cost-effectively (if your best guess is that increasing/decreasing agricultural land increases welfare)? I think interventions targeting invertebrates are very unlikely to be the most cost-effective if moral weight is proportional to "individual number of neurons"^n (model 1) or "basal metabolic rate (BMR) at 25 ÂşC"^m (model 2). Below is a graph showing the most cost-effective intervention among 16 I analysed as a function of n and m. The Shrimp Welfare Project's (SWP's) Humane Slaughter Initiative (HSI) only has the highest cost-effectiveness for values of m slighly below 1.
People endorsing deontology often care about not causing harm, and I believe corporate campaigns for cage-free layers may increase suffering a lot. I estimate their cost-effectiveness is -7.33 kQALY/$ for n = 0.2, and n = 0.188 explains very well the sentience-adjusted welfare ranges presented in Bob's book (see the last graph in my comment above).
Why would we have a special duty to help chickens, but not soil ants and termites?
I think reasonable moral parliaments will give some weigth to increasing suffering a lot being bad, and I believe corporate campaigns for cage-free layers may increase suffering a lot. So I think corporate campaigns for cage-free layers may easily be bad under many moral views.
On special duties, my deontological intuition is more about caring more about righting wrongs of cruelty or exploitation than wrongs of incident. For example, if I am angry at Alice and punch her, and she falls over onto Bob who also gets hurt, both deserve compensation. But if I had to choose who should receive compensation first, I’d choose Alice. I feel a special duty towards her because I directly and maliciously impacted her welfare. This intuition can be subverted if the difference in cost-effectiveness is big enough—if I punch Alice and she tumbles onto Bob who receives a life-threatening injury, I care more about rectifying Bob’s injury than the punch to Alice.
On my invertebrate giving, it’s been SWP but also the Insect Institute and Arthropoda. On your plots, I’d need to understand much more about the vectors for welfare you’re assuming for each intervention. One potentially relevant difference is that I put more probability mass on models like the undiluted experience model, or others where the neuron count exponent is very low. However I’m deeply uncertain about moral weights in general. It’d also be helpful for me to see the top 5 interventions (instead of just the top one) as a function of the exponent, since otherwise I don’t have visibility into how the rank order moves as the exponent changes.
Would you still feel a special duty if you expected your action to impact Alice as much as Bob? For example, you can assume someone else will impact Bob in exactly the same way as you impact Alice if you chose to impact Alice, and that you are aware of this.
Do you consider you have "directly and maliciously" impacted chickens? I do not think this follows from eating eggs or chicken.
The most cost-effective interventions for n close to 0 are the ones which most cost-effectively increase the welfare of soil invertebrates, and these are the ones that most cost-effectively increase or decrease agricultural land, not ones targeting farmed invertebrates.Â
I have a graph with this in my analysis. It is at the start of the section "Cost-effectiveness per dollar". Below is the cost-effectiveness as a function of n of the interventions which are the most cost-effective for at least one value of n, and HSI. The Good Food Institute (GFI) is in green, HSI in blue, GWWC in red, and Helen Keller Intl's (HKI's) vitamin A supplementation in purple. For n = 0, I estimate GFI is 982 (= 2.18*10^6/(2.22*10^3)) times as cost-effective as HSI assuming decreasing agricultural land increases animal welfare. Flipping the signs of the welfare of soil invertebrates, GFI and HSI would decrease welfare, and GWWC and HKI's vitamin A supplementation would increase welfare.
The conclusion that HSI is not the most cost-effective intervention for n = 0 is robust because it is far from the one which increases/decreases agricultural land the most cost-effectively.
Hey Vasco, I have some points:
I see no reason why your conclusions might be wrong, but they are still fairly weird, even to many EAs like myself. This increases my moral uncertainty about the topic.
Hi PAMC.
Funding Rethink Priorities' (RP's) research on moral weights led by Bob Fischer, which can be done with a restricted donation to RP's worldview investigations team (WIT).
I prefer model 2 because it accounts for effects on soil microorganisms. Funding The Navigation Fund's (TNF's) cage-free campaigns fund has the highest cost-effectiveness for m equal to 0.85, and between 1 and 1.85 (among multiples of 0.05), but I can easily see m taking values below 0.85. I am very uncertain about moral weights.
I analysed concrete funding opportunities (like donating to GWWC) because cost-effectiveness may vary significantly within the same type of intervention or area.
I think interventions targeting farmed chickens will overwhelmingly build capacity to help farmed vertebrates, including via motivating the general population to help farmed vertebrates. I do not know whether interventions helping farmed vertebrates increase or decrease welfare. So I also do not know if the capacity built via helping farmed chickens will increase or decrease welfare.
In any case, it would be a surprising and suspicious convergence if helping farmed chickens was the best way to build capacity to help soil invertebrates. Here are some of my related thoughts.
I believe funding Arthropoda Foundation is the most cost-effective way of building capacity to help soil invertebrates. Here are some related thoughts from Bob.
Thanks Vasco,
I don't think "helping farmed chickens" is the best way, but I do think that the foot-in-the-door strategy proposed by Tobias Lenaert is (among?) the most effective strategy if the goal is to change other people's minds. This means requesting very small changes, because in humans, values often follow behaviour. Thus, I believe political campaigns can help expand the moral circle, and I do not see a way to arrive at non-vertebrates without going through chickens (and fish, and other farmed animals in general).
You may be interested in the article "Beyond the Expanding Circle: Towards a Pluralist Account of Moral Progress" by Rafael Ruiz. Here is a crosspost for a summary by Rafael.
Do you think supporting corporate or political campaigns for cage-free layers is close to the most cost-effective ways of helping invertebrates? I believe the most cost-effective ways of helping any given group of animals are usually optimised with such animals in mind. For example, I would target (optimise for increasing the welfare of):
Here are related thoughts from Lewis Bollard.
Here are related thoughts from Aidan Kankyoku.
I think the crux is sufficiency vs necessity. I think for people to care about non-vertebrates, they also need to care about factory-farmed animals, because in the vast majority of people's minds, it'd be ridiculous to care about ants but not chickens. Perhaps you think that it's the same amount of effort to convince them to care about (ants AND chickens) as just chickens, but I'd guess the opposite. I believe it is easier to do it in steps: first vertebrates and then non-vertebrates. Part of the reason is that suffering is much more visually apparent in vertebrates.
I don't think any of that contradicts the idea that caring about domestic animals will not carry over to farmed animals. If you tried to apply my argument to this case, I think you'd get a different diagnosis: (i) people care about cats and dogs, (ii) but believe the suffering of cats and dogs is not that important compared to other issues (world hunger and the like), and additionally it's not something they are personally involved in.
So, all in all, I don't find the arguments above convincing enough to convince people to care about nematodes.
What do you think this concretely implies? I may agree that the most cost-effective way to get 10 people to care about ants flows through getting 10 people to care about vertebrates. However, it does not follow that supporting corporate or political campaigns for cage-free layers is close to the most cost-effective ways of helping invertebrates today.
Which people? I think research on the moral weight of nematodes is valuable.