I see myself as a generalist quantitative researcher.
Vasco Grilođ¸
Karma: 7,250
Thanks for sharing, RĂan. People who prefer audio may also be interested in the podcast Introduction to Utilitarianism, which covers the content on the website.
This would only be 100 % correct if the welfare per time of the practically maximally happy life as a fraction of the welfare range is constant across species.
In this case, âmaximum welfare of a chicken-yearâ/â(âmaximum welfare of a chicken-yearâââminimum welfare of a chicken-yearâ) = âmaximum welfare of a human-yearâ/â(âmaximum welfare of a human-yearâââminimum welfare of a human-yearâ) â âmaximum welfare of a chicken-yearâ = (âmaximum welfare of a chicken-yearâââminimum welfare of a chicken-yearâ)/â(âmaximum welfare of a human-yearâââminimum welfare of a human-yearâ)*âmaximum welfare of a human-yearâ. Since âwelfare range of chickensâ = (âmaximum welfare of a chicken-yearâââminimum welfare of a chicken-yearâ)/â(âmaximum welfare of a human-yearâââminimum welfare of a human-yearâ), âmaximum welfare of a chicken-yearâ = â1 AQALY in chickensâ, and âmaximum welfare of a human-yearâ = â1 QALYâ, â1 AQALY in chickensâ = âwelfare range of chickensâ*â1 QALYâ. So, given the condition I mentioned in the footnote, one can get the welfare in QALYs mutiplying the welfare in AQALYs by the welfare range.
Thanks for the great post, Omnizoid!
This makes them around 30 times better at reducing suffering and promoting well-being than the highly effective animal charities focused on chicken welfare which themselves are hundreds or thousands of times more effective than the best charities helping humans.
I have updated my cost-effectiveness analysis of corportate campaigns for chicken welfare. Now I estimate the Shrimp Welfare Project is 278 and 117 times as cost-effective as broiler welfare and cage-free campaigns (instead of âaround 30 times betterâ), and that these are 156 and 371 times as cost-effective as GiveWellâs top charities (i.e. more like âhundredsâ instead of âhundreds to thousandsâ of times more cost-effective than the best charities helping humans).
Thanks, Aidan. For reference, I estimated corporate campaigns for chicken welfare are 25.1 times as cost-effective as School Plates, which is a program aiming to increase the consumption of plant-based foods at schools and universities in the United Kingdom.
Thanks for the comment, Chris!
Am I thinking about this correctly?
Yes.
And if so, is there a good resource for actual welfare values for farmed animals, rather than the theoretical ranges?
I have estimated the welfare per living time of chickens in various conditions in animal quality-adjusted life years (AQALYs) per chicken-year. 1 AQALY corresponds to 1 year of a practically maximally happy life. As a rough approximation, you can get the welfare in QALYs mutiplying the welfare in AQALYs by Rethink Prioritiesâ median welfare ranges[1].
Animal Broiler in a conventional scenario Broiler in a reformed scenario Hen in a conventional cage Hen in a cage-free aviary Welfare per living time (AQALY/âyear) -2.27 -0.161 -1.69 -0.333 I have some estimates for shrimp too (this post has estimates for chickens, but these rely on underestimates of the time they spend in pain, whereas the ones above try to correct for this).
- ^
This would only be 100 % correct if the welfare per time of the practically maximally happy life as a fraction of the welfare range is constant across species.
- ^
Thanks, Karolina! Do you have any thoughts on whether it is better to donate to the AWF or Shrimp Welfare Project? I estimate this is 278 and 117 times as cost-effective as broiler welfare and cage-free campaigns.
Thanks for clarifying, Daniela! Keep up the good work.
I plan to post an updated cost-effectiveness analysis of chicken welfare campaigns in 1 or 2 weeks.
I very much agree with your decision of not recommending ACEâs RCF, and I think your evaluation was great.
Another example where we think ACE could have engaged more robustly with their CEAs is in cases where ACE did not reach estimates for Suffering-Adjusted Days (SADs) averted per dollar.
Did you have the chance to look into the pain intensities Ambitious Impact uses to calculate SADs? They are not public, but you can ask Vicky Cox for the sheet. I think they hugely underweight excruciating pain, such that interventions addressing this (like Shrimp Welfare Projectâs Humane Slaughter Initiative) have their cost-effectiveness underestimated a lot. Feel free to ask Vicky for my comments on the pain intensities.
Nitpick. ACEâs additional funds discount for low uncertainty should be higher than 0 % (and lower than 20 %), or there is a typo in the table below?
ACE considers charitiesâ priorities for additional funds â in other words, how the charity would use unexpected additional funding â and assesses âuncertainty about effectiveness and tractability of plansâ â how likely they think each line item in the charityâs plan is to experience diminishing returns or not play out as expected under the charityâs growth plan. ACE assesses each line item in a charityâs proposed future plans on a scale from very low uncertainty to very high uncertainty, with a corresponding discount applied to the cost of the line item.
Category Discount Very low uncertainty 0% Low uncertainty 0% Moderate/âlow uncertainty 20% Moderate uncertainty 40% Moderate/âhigh uncertainty 60% High uncertainty 80% Very high uncertainty 100%
Welcome to the EA Forum, nailjim!
The website links to the FAQ for info about the calculation method, but I didnât see the info there.
Did you mean to link to Starbucks?
From your 2024 evaluator research page:
Note: we were highly time-constrained in completing this second iteration of our project. On average, we spent about 10 researcher-days per organisation we evaluated (in other words, one researcher spending two full work weeks), of which only a limited part could go into direct evaluation work â a lot of time needed to go into planning and scoping, discussion on the project as a whole, communications with evaluators, and so on.
I think you did great based on the time you spent. Going through your evaluations, I thought to myself âgood pointâ (or similar) so many times! Have you considered fundraising specifically for your evaluator research? I would be interested in supporting you as long as some (e.g. 1â3) of the additional funds go towards evaluating animal welfare evaluators.
From your evaluation of ACE MG:
Using quite an expansive definition of institutional diet change advocacy, we identified that two of the 30 EA AWF grants (constituting 0.27% of funding) went to these kinds of activities, compared to six out of 28 MG grants (constituting 11.12% of funding).
[...]
Are institutional diet change advocacy campaigns competitive with other opportunities?
To understand whether this difference constituted a concern, we asked experts for their views on the promisingness of institutional diet change advocacy.
[...]
On existing evidence, some institutional diet change advocacy campaigns can meet a cost-effectiveness bar, but this is not typical.
Itâs probable that most of the impact of institutional diet change advocacy runs through direct impact rather than movement building.
Which diet change interventions are more cost-effective than cage-free corporate campaigns? Meaningfully reducing meat consumption is an unsolved problem, and I suspect many diet change interventions are harmful due to leading to the replacement of beef and pork with poultry meat, eggs (which can be part of a vegetarian diet), fish and other seafood.
In any case, I do not think this is that important for your recommendation. You found only 11.1 % of the money granted by ACE MG went to diet change interventions, and it does not look like marginal grants were super related to diet change.
In support of being more funding constrained, ACE also provided us with a summary of how they would likely have disbursed an additional $200Kâ300K USD in the 2024 round. This included a mixture of fully funding more of the opportunities that they only partially funded, and funding some opportunities that missed out altogether in the round â in particular, two opportunities related to insect farming. We attempted to naively model the scope of some of the opportunities that MG indicated they would have funded and concluded that we expect these opportunities to have relatively high marginal cost-effectiveness â competitive with other opportunities ACE MG funded.
However, because we expect the approach of grantmakers/âevaluators to change more slowly than the work/âprograms of an individual charity, we are not comfortable relying on the information we used last year without undertaking a reinvestigation,[1] which we donât expect to do (see below)
I think the rate of change is overwhelmingly a function of size, not of whether it respects grantmakers/âevaluators or work/âprograms. I would expect THLâs work/âprogram to change more slowly than the Animal Welfare Fund. In any case, I agree with your decision of not recommending THL. As you say, it is more consistent with the scope of your project:
We generally base our charity and fund recommendations on our evaluating evaluators project, for reasons explained here. This year, the evaluators and grantmakers weâll rely on â based on our evaluations â will be EA Fundsâ Animal Welfare Fund and ACEâs Movement Grants, and neither of these currently recommends THL as a charity (as neither makes charity recommendations).
Thanks, Aidan and Sjir! I really like this project.
While our decision is not to recommend FP GHDF at this time, we would like to emphasise that we did not conclude that the marginal cost-effectiveness of the GHDF is unambiguously not competitive with GiveWellâs recommended charities and funds â in fact, we think the GHDF might be competitive with GiveWell now or in the near future.
I do not understand why you âthink the GHDF might be competitive with GiveWell nowâ. From your report, it seems quite clear to me that donating to GiveWellâs funds is better now. You âlooked into [3] grants that were relatively large (by FP GHDF standards), which we expect to be higher effort/âqualityâ, and found major errors in 2 grants, and a major methodological oversight in the other.
Grant 1
When the grant was made, the opportunity was determined to be 42x GiveDirectly. However, on a re-evaluation for a later grant to the same organisation (where the original BOTEC was re-purposed), an extra zero was discovered in one of the inputs that had been used â which, when remedied, reduced the overall estimate to 12x GiveDirectly.
[...]
Grant 2
In the second grant we looked into, the effect size used to estimate the impact of the grant applied to children above a certain age, but annualised total childhood mortality in the region was used as the assumed base mortality rate to which the effect size was applied. Because most childhood mortality occurs in the first few months of life, we believe this is a mistake that overestimated the impact of the intervention by about 50%. When we applied adjustments to the model to account for this, the overall cost-effectiveness of the BOTEC fell below 10x GiveDirectly.
[...]
Grant 3
In the third grant we reviewed, our major concern was that FP modelled the probability that a novel technology would be significantly scaled up at 35%, with what we considered to be insufficient justification. This was particularly concerning because small reductions to this estimate resulted in the BOTEC going below 10x GiveDirectly.
Notably, in a subsequent evaluation conducted about 6 months after the first, FP adjusted this probability down to 25%. While we think it is positive FP re-evaluated this:
1) We still think this probability could be too high (for similar reasons to those noted in Grant 1).
2) Plugging this number into the original BOTEC sends the estimated cost-effectiveness below 10x GiveDirectly, which reinforces the point that the original BOTEC should not have been relied on in the original grant without further justification.
3) In the more recent evaluation â which repurposed the original grant evaluation and BOTEC, but where the same probability was modelled at 25% rather than 35% â there was no reference to why the number was updated or additional information about why the estimate was chosen. This increases our concerns around FPâs transparency in justifying the inputs used in their BOTECs.
Thanks for the post, Michael. Thanks to the EA Forum team for sharing it in the EA Forum Digest.
Research advice from Carl Shulman (doc created by Claire Zabel in 2016, according to this page)
I think the advice Carl gives in sections âGood habitsâ, âResearch project processâ and âMindsetâ is quite good.
Thanks for sharing! Great work.
Likewise, we encourage researchers to think clearly about the difference between reducing all MAP consumption and reducing just some particular category of it. RPM is of special concern for its environmental and health consequences, but if you care about animal welfare, a society-wide switch from beef to chicken is probably a disaster.
I concluded the harm caused to humans by the annual GHG emissions of a random person is 0.0660 DALY, and that caused to farmed animals by their annual food consumption is 4.04 DALY, i.e. 61.2 times as much. In my mind, this implies one should overwhelmingly focus on minimising animal suffering in the context of food consumption.
Thanks for sharing, Camille! Relatedly, people may like James Ăzdenâs post The default trajectory for animal welfare means vastly more suffering. For somewhat opposite evidence, there is my post Farmed animals may have positive lives now or in a few decades?, and Robert Yamanâs How to Be a Techno-Optimist for Animals.
Thanks for sharing.
I think it is interesting that this is the only reference to cost-effectiveness estimates among so many recommendations. The case for following your recommendations seems quite unclear to me absent such quantification. However, I am glad you still shared them.