PhDs in physics (thermodynamics of ecosystems), moral philosophy (animal rights) and economics (altruistic motivation and incentives for blood donation), co-founder of EA Belgium, environmental footprint analyst at Ecolife
Stijn Bruers 🔸
Karma: 899
Hi Vasco,
Perhaps I understand your position. Here is a line of argument that results in your conclusion based on total utilitarianism. Suppose for simplicity every individual can choose between four options:
A: veganism: not buying and eating animal products. This option generates the second lowest total welfare, higher than option B.
B: unhappy animal farming: buying and eating products from factory farms where animals live unhappy lives. This generates the lowest total welfare.
C: happy animal farming: buying and eating products from farms where animals live happy lives. This option generates the second highest total welfare, lower than option D but higher than A and B.
D: animal sanctuary: eating vegan and donating money to a sanctuary that breeds animals and give those animals long and very happy lives. This generates the highest welfare.
In terms of total welfare, we have D > C > A > B. Therefore, if everyone was total utilitarian, everyone should choose D. In extremis, total utilitarianism entails that everyone should sacrifice themselves a lot in order to bring into existence huge numbers of very happy animals that live on a sanctuary. However, if only you are total utilitarian, then according to act utilitarianism it would be wrong for you to choose D when other, non-utilitarian people are extremely reluctant to choose D. Sidenote: if you were a rule utilitarian instead of an act utilitarian, you could have to follow the rule that everyone has to follow, namely choosing option D, even if other people do not follow that rule (i.e. do not choose that option D). So I assume you are an act utilitarian.
There are two things to consider: which option you should choose for yourself, and which option you should choose for other people. I’ll argue that you should choose option A for yourself and, importantly, option C for others.
As a total act utilitarian, you should choose option A for yourself. Sure, option A is dominated by options D and C. But those two options are more expensive: in C you have to pay extra money for the meat, because meat from happy animals is more expensive than vegan products and meat from unhappy factory farmed animals. In option D you have to donate money to the sanctuary. However, it is very likely that donating your money to the most cost-effective animal welfare charities improves total welfare more than donating that money to an animal sanctuary or a farm where animals are happy. In particular, the extra money that you would spend on high-welfare meat when you choose option C, could better be spend on effective campaigns to transition people from option B to C.
This brings me to the second issue: which option should you choose for other people? Most people currently choose option B. Persuading them to choose option C might be feasible with campaigns that require a bit of money, whereas shifting people to option D might be unfeasible, no matter how much money you throw at such a campaign. At first sight, total utilitarianism seems to suggest that the best strategy is to have people switch to option D, but that may be unfeasible or too costly. Note that in terms of total welfare, it is better to persuade people to choose option C instead of option A.
So there you have it: choose veganism for yourself and happy meat for others. That would be the conclusion of a total act utilitarian.
I meant that prohibiting abortion violates the right to bodily autonomy of the mother: her body is required for the pregnancy, and the pregnancy was against her will. So the body of the mother is used as merely a means. Abortion and killing a foetus does not violate the right to bodily autonomy of the foetus, because the body of the foetus is not used as a means to achieve someone else’s ends. If the body of the foetus did not exist, there was no pregnancy and then the objective of the mother, not to be pregnant, is automatically achieved. So the presence of the body of the foetus is not necessary. So the body of the foetus is not used as a means.
Those 36% of US population that are against abortion, do not count, because they have an inconsistent ethic that involves unwanted arbitrariness (discrimination). Their ethic is speciesist, whereby human foetuses have a higher moral status than non-human animals. They believe that foetuses should not be killed whereas non-human animals are allowed to be killed. Such speciesist discrimination is arbitrary, and cannot be wanted by non-human animals. Hence unwanted arbitrariness. If such unwanted arbitrariness would be permissible, then we would be permitted to arbitrarily exclude the opinions of those people that are against abortion and arbitrarily exclude those people from democracy.
How many people would be against killing and eating individuals if those individuals had positive lives? If speciesism is excluded, and hence ‘individuals’ can refer to humans as well, then I expect a vast majority would be against such killing. Even if that means the individuals are not brought into existence.
About your total utilitarianism: the loss of welfare of a human who loves the taste of meat but is no longer allowed to eat meat, is smaller than the loss of welfare of all the animals who would be prematurely killed and eaten by that human. Hence, breeding animals but not killing and eating them, would increase total welfare. Also: the loss of welfare of a human who gives a bit of money to an animal sanctuary is smaller than the loss of welfare of animals on a sanctuary that lacks money to buy things for the animals. Hence, to increase total welfare, the human should give a bit of money to the animal sanctuary. And by the same reasoning he should give some money to another sanctuary that breeds and helps extra animals. And so, according to total utilitarianism, the human is not allowed to eat meat and is obligated to donate a lot of money to animal sanctuaries.
You are correct to assume that people would not want to farm lots of animals with super high welfare if they were not being raised for meat. That means those people are not total utilitarians.
If you want to maximize the ratio of annual welfare and cost, and if annual welfare does not explicitly depend on the lifespan of the animals, then you run into the replacement problem: the belief that killing an animal painlessly and bringing another animal into existence that has the same momentaneous welfare, is morally neutral. That would assume that animals would not have a personal identity over time. But there is evidence that animals are concerned about their future and do not want to be prematurely killed.
To be honest, I think that you are deluding yourself by believing that expectational total hedonistic utilitarianism is not that demanding. The belief that huge sacrifices would cause you to be less effective in doing good, seems to be like merely a rationalization. I don’t know of any total utilitarian who could not do a little more sacrifice in order to do more good. You can donate to the point that you are very poor and still work as hard, or even harder, to earn as much money, or even more money. I think total utilitarians should at least be honest and acknowledge that they are not doing all the things that would maximize total welfare. All total utilitarians that I know, are able to donate a little bit more on the margin and do more good.
About future human-years: a human who exists today can validly complain today that his/her welfare in the future is discounted. A human who does not exist today and will never come into existence cannot validly complain today that his/her welfare in the worlds where s/he is brought into existence in the future, is discounted. Hence, future human-years of humans who exist today always count at least as much as future human-years of humans who will never exist. Note that most humans believe that they have a personal identity over time, that they do identify themselves with future human-years.
Most people have the moral intuition that the right to bodily autonomy trumps the right to live. We see that in dozens of moral dilemma’s, so it is pretty coherent. That right to bodily autonomy is also special in many other ways, and it can be directly derived from the moral theory of mild welfarism, as the right to discount the welfare of others if those discounted individuals cannot validly complain. The question is how much does that right to bodily autonomy trump other rights? In terms of mild welfarism: how much welfare are we allowed to discount? I suggest for collective decisions, we have to democratically decide how much we are allowed to discount, or how strong the right to bodily autonomy is.
If you endorse expectational total hedonistic utilitarianism, you should be against happy animal farming because those animals would be much better-off on a sanctuary where they live much longer with a higher welfare. Same goes for happy slavery: those happy human slaves would be better-off if they were not treated as slaves. According to total utilitarianism, you may have a duty to sacrifice yourself by bringing into existence huge numbers of (slightly) happy beings and taking care of them. Basically the very repugnant conclusion. That utilitarianism is rather totalitarian utilitarianism: extremely demanding. To avoid this repugnant conclusion, and hence to justify happy animal farming, you should believe that welfare is not interpersonally comparable, i.e. that you cannot sum everyone’s welfare. But you believe that the welfare of a shrimp is comparable to that of a human.
The fundamental difference between future human-years of existing and future people is based on the difference between existing and non-existing people. Existing people can complain against their welfare being discounted in those states where they exist, but possible future people who are never brought into existence, will never exist and hence cannot complain against their welfare being discounted in those states where they would exist. Hence, I don’t think this difference based on existence is arbitrary.
I’m not against abortion (except very late term abortion, to some degree), because unwanted pregnancy violates a woman’s right to bodily autonomy (that trumps the right to live, to some degree). During such pregnancy, the body of the woman is used as a means against her will. That is the difference between abortion and killing a baby after birth. In terms of my mild welfarist ethic: the woman has the right to discount the welfare of the foetus, only during pregnancy, when the body of the woman is used as a means to keep the foetus alive. That right to discount welfare entails the right to bodily autonomy. Rejecting this right would mean that we have to make huge sacrifices, no utilitarian is willing to make.
I wonder, if you say happy animal farming is permissible, what do you think about happy slavery (breeding people to be used as slaves, guaranteeing a net-positive welfare) or happy human cannibalism?
Future human-years can only be discounted (in this context) if that discounting results in choosing the state where those humans do not exist. So, future human-years of humans who already exist, cannot be discounted. If there is a non-zero personal identity over time, then your future self and your current self are the same person and it will be impossible for me to choose the state where that person does not exist.
I believe that kids having a preference not to be slaughtered, is a very good reason not to have kids, when someone decides to have happy kids in order to slaughter and eat them. Kids having a preference to live longer is not a good reason not to have kids. Note a difference here between slaughtering and not living longer: it is possible not to slaughter a kid, but it may be impossible to have that kid live as long as it wants.
In my opinion, we are allowed to consider the lives of possible people as neutral if that consideration results in those possible people not existing. In other words: we are allowed to discount the welfare of possible people, if the situation with the highest sum of discounted welfare is one where those welfare-discounted people do not exist. The welfare-discounted people cannot complain against their welfare being discounted, because in the chosen optimal situation, those people do not exist. This idea creates the procreation asymmetry: it is ok not to cause the existence of a happy person, but not ok to cause the existence of a miserable person (all else equal). See my theory of mild welfarism: https://stijnbruers.wordpress.com/2022/08/23/mild-welfarism-avoiding-the-demandingness-of-totalitarian-welfarism/ and the asymmetry: https://stijnbruers.wordpress.com/2024/06/17/the-asymmetry-in-population-ethics-and-deontological-ethics/
My theory of mild welfarism gives two reasons why breeding happy farmed animals to kill and eat them, or breeding happy slaves, or giving birth to happy kids to slaughter and eat them, are not permissible. Those things violate the rights of bodily autonomy and procreation autonomy.
I agree with maximizing the additional welfare per $ spent. but I’m skeptical that turning the agricultural land into an ecosystem with many arthropods, or keeping the agricultural land, requires less dollar spending for extra welfare, compared to turning that land into wild grassland with horses. First, the welfare range of a large herbivore is larger than that of an arthropod, second, the fertility rate of a large herbivore is lower, and third, its lifespan is longer. Plus, that grassland with wild horses can have flowers that attract bees and horse manure that attract dung beetles. Of all arthropods, I believe bees and dung beetles are the best in terms of maximizing global welfare. Those insects are most likely to have a positive welfare and a higher welfare range, and are less harmful to other animals, compared to the other arthropods. A grassland can also have more bees and dung beetles than agricultural cropland used for animal feed. Improving the welfare of animals on agricultural land or in other ecosystems is likely more costly than improving the welfare of animals on that wild grassland. https://stijnbruers.wordpress.com/2025/03/11/how-to-systematically-reduce-wild-animal-suffering-in-the-near-future/
I agree with taking that perspective of the animals. However, in this context it’s complicated: if the animal does not exist, it has no perspective and also no preference for existence with a positive welfare. And if you bring an animal into existence, give it a positive live, use that animal for food and slaughter it, it is likely that from that animal’s perspective, that animal would prefer a longer, more positive live at an animal sanctuary. So, if the conditions were really equivalent to those of healthy animals in sanctuaries, or pets, that means no slaughtering. You can eat sanctuary animals that died naturally of old age, but not happy farmed animals that are slaughtered.
Arthropods tend to have higher reproduction rates (more offspring per adult) than large herbivores, and I believe high reproduction rates negatively correlate with welfare of the offspring. Also: as the reproduction rate of most farmed animals is higher than those of wild large herbivores, replacing farmed animals by wild large herbivores (when agricultural land is replaced by grasslands with wild horses) would increase animal welfare and decrease suffering.
There are several (strong, according to many people) moral objections to have happy animal farming, where farmed animals with a net-positive welfare are slaughtered and eaten.
https://stijnbruers.wordpress.com/2016/11/05/can-we-eat-happy-meat/
https://philpapers.org/rec/BRUPEA-5
As for the effects on wild animals, I guess many of the agricultural land that is now used for animal farming (including feed crops), can best be turned into grasslands with large herbivores (wild horses,...). Those types of ecosystems may have lower animal suffering densities (amount of suffering per square mile) than animal farming.
A few years ago, I made a back-of-the-envelope cost-effectiveness calculation for cultivated meat R&D. https://stijnbruers.wordpress.com/2020/08/10/the-extreme-cost-effectiveness-of-cell-based-meat-rd/
That calculation assumed a tractability of 10%, which can be interpreted as: cultivated meat will enter the market sooner or later, and when it does, it will capture 10% of the animal-based meat market, i.e. replace 10% of animal-based meat by cultivated meat. That 10% represents the animal-based meat that will not be substituted by plant-based meat, but will be substituted by cultivated meat. This is in line with the Bryant-Sanctorum study that you mentioned. Even with this low tractability, corresponding to a small complementarity between plant-based and cultivated meat, funding acceleration in cultivated meat R&D remains highly effective. The reason is its neglectedness: in terms of investments and funding, cultivated meat is roughly 10 times more neglected than plant-based meat. If you are right, and plant-based meat will capture more of the animal-based meat market than cultivated meat, it would mean that the current global allocation of R&D funding between plant-based and cultivated meat seems about right. You may believe ten times more in plant-based meat than in cultivated meat innovation, which means you may allocate more (but not all!) funding to plant-based meat.
My major concern about cultivated meat innovation cost-effectiveness, is the longer timeline. The types of cultivated meat that could replace animal-based meat and could not be replaced by plant-based meat, are whole tissue cultivated meat, and those will probably take a few decades before development and becoming competitive with animal-based whole tissue meat. It is likely that within a few decades, we will have transformative artificial intelligence (TAI) that could do research 10 times faster. If we have such TAI in say 2035, it could do all the research of the next ten years in only one year. If that happens, we could simply not invest in (whole tissue) cultivated meat R&D for ten years, wait until 2035 and then let the TAI do all that research in one year. Instead of having the results by the end of 2035, we have them by the end of 2036, i.e. 11 instead of 10 years later. That’s only a 10% longer timespan. Hence, this consideration would make the cost-effectiveness of cultivated meat R&D 10 times lower. I’m a bit skeptical that TAI will be able to do all cultivated meat research 10 times faster, because much of that research requires lab experimenting with living tissues in bioreactors, and such research cannot easily be accelerated by computers and robots. But who knows: perhaps TAI can quickly calculate the ideal bioreactor design and the optimal cell types and growth medium to use.
The difference in cost-effectiveness estimates of animal welfare reforms versus alt protein is only a factor of three. Given the large uncertainties in these estimates, which can easily result in uncertainty ranges of a factor of 10 or more, my rule of thumb is that if the first best cost-effectiveness estimates do not show a difference larger than an order of 10, I assume the difference in cost-effectiveness is not (statistically) significant.
The probability that animal welfare reforms is more cost effective than alt protein may be higher than 50%, because you easily derived such a result with reasonable assumptions, but given these large uncertainties, I expect the probability is not much larger. Say 51% likelihood that animal welfare reforms are more effective than alt protein. If you keep the estimates to yourself (i.e. you don’t make recommendations to others based on your estimates), and if you are a small donor, there are no decreasing marginal returns and no coordination issues, and hence it is rational to donate 100% of your own donations to the most effective charity according to your first best estimates, which has a 51% probability of being the best charity. However, for large donations, or for the total donations of the EA community as a whole, there may be decreasing marginal returns, and I think in that case diversifying is more important than putting all eggs in one basket (pun intended) with a 51% probability of being the best basket. I believe as a rule of thumb, if you communicate your results to other EA people and give recommendations, you have to consider the total donations of the EA community instead of the donations of an individual, small EA donor. Otherwise your recommendations should mention the need for coordination with other donors, but that message is more complex. So you can recommend: “Donate everything to this charity that has a 51% probability of being the most cost-effective, but keep an eye on what other effective altruists are doing, how much they are donating to the different charities.”, or you can recommend “Diversify: donate something to animal welfare reforms and the rest to alt protein.”
Hence, given you published the above post and made it public to the community, I suggest you recommend diversifying donations to both animal welfare reforms and alt protein. And privately you can donate 100% of your donations to what you believe is the most effective charity, as long as you do not recommend others to do the same.
on a first view, I think the cost-effectiveness of egg replacement precision fermentation R&D is the same order of magnitude as my (low) estimate of cultivated chicken meat R&D. Your precision fermentation estimate: roughly 1 layer hen life spared per dollar. My cultivated meat estimate: roughly 10 broiler chicken lives spared per dollar. But lifespan of layer hen is ten times longer than broiler chicken, and suffering intensity is about the same, so in terms of avoiding suffering, both are comparable in cost-effectiveness
I’m curious how your calculation compares to my model, described here: https://stijnbruers.wordpress.com/2020/08/10/the-extreme-cost-effectiveness-of-cell-based-meat-rd/
[Question] Why do EA longtermists not discuss the alleged problem of cognitive decline?
small herbivorous animals with high reproduction rates compete for food with large herbivores. A large herbivore captures more of the plant biomass than a small animal, and consequently large herbivores prevent large populations of small herbivores.
Good question: it is unsure. Problem with large animals is that they more likely accidentally kill (trample) small insects. That might be a good thing is the reproduction rates of insects are so high that they have net-negative welfare levels. The large herbivore reduces the population sizes of the insects that have lives not worth living. An advantage of the large animals in grasslands, is that their manure attracts dung beetles, and the flowers in the grassland attract bees. Now, dung beetles and bees are the best, kindest or nicest kinds of insects: they do not hunt and kill other insects, they are not parasites, they do not compete much for food with other animals, they fertilize the soil and the plants which means more food for other animals, they are highly intelligent and sentient. If you want to help insects, I’d say prioritize helping bees and dung beetles.
Good question! I have no idea.
A problem with broader conservationist initiatives is that they more likely support interventions that could increase wild animal suffering, e.g. by helping predator species or species with high reproduction rates.
How to systematically reduce wild animal suffering in the near future
Rational Animal Ethics (my top 10 ideas of all time)
I agree, that’s why I generally support it.
I generally support this idea of diet offsetting, although purely morally speaking I have several objections, explained here: https://stijnbruers.wordpress.com/2019/08/22/carbon-offsetting-versus-meat-offsetting/
There are morally relevant differences between carbon offsetting and meat offsetting.
Nice work, Soemano. Strongly upvoted.
In my calculation, which you refer to (Bruers, 2024), I used the smaller own-price elasticities and higher cross-price elasticities of demand (based on Lusk, J. L., & Tonsor, G. T. (2016). How meat demand elasticities vary with price, income, and product category. Applied Economic Perspectives and Policy, 38(4): 673-711.), and perfect elasticity of supply. When I use your model, but using the Lusk & Tonsor elasticities of demand, I obtain an increase in chicken meat and pork, as in my original calculation. I agree that you used better estimates of own and cross price elasticities of demand (from a meta analysis, and from 2024 paper), so I have more confidence in your results.
When I did your calculations, I noticed a small typo in your paper, ‘Table 2, model input data’: I think L_d and L_s for chicken and pork should be reversed (your values for chicken correspond with my values for pork).
I expect animal welfare in D to be higher than in C, almost by definition. In D, the purpose is to have as many animals as happy as possible, without killing them. Whereas in C, the prime objective remains killing and eating the animals. The farmers do not intend to maximize the welfare of the farmed animals. The population in D may be lower than in C, but the population size in D is chosen to maximize animal welfare, whereas the population size in C is chosen to maximize farmers profits. If population size in C is larger than in D, the welfare of the animals in C is suboptimal.
With choosing option C for others, I actually meant not trying to reject option C for others, e.g. not campaigning against option C, not spending money on trying to switch people to option A instead of option C. I agree that the most cost-effective invertebrate welfare interventions are more effective than interventions to switch people from B to C. Or, if you want, you can add another option B’, which is similar to B but with reduced animal suffering. Your favorite invertebrate welfare interventions correspond to switching people from B to B’, and that might at the moment be more effective than switching them from B to C.
Thanks for the conversation!