This anonymous essay was submitted to Open Philanthropy's Cause Exploration Prizes contest and posted to the Forum with the author's permission.
If you're seeing this in summer 2022, we'll be posting many submissions in a short period. If you want to stop seeing them so often, apply a filter for the appropriate tag!
Of the approximate $200 million in funding aimed at improving the lives of farmed animals, only about 3% is spent on research and measurement and evaluation. This seems remarkably low, given how little the animal advocacy movement knows about which interventions are actually effective in changing behaviours, influencing policy or affecting animal product production. My very naive BOTEC puts the leverage factor of improved research at about 12x, meaning I believe $2 million of animal advocacy R&D could influence $23 million (2% of the next five years of funding).
Funding intervention research early-on can improve the trajectory of the animal advocacy movement, and counterfactually improve the number of animals we help over the long-term. Not only does it provide a large value of information, I believe it is fairly tractable. For example, one can:
- Increase the pipeline of talented researchers by funding summer research fellowships
- Support existing academic research groups by academic network building, providing administrative support and hiring, as done by BERI.
- Release requests for proposals on large-scale and neglected advocacy methods.
- Incubate a new charity to support animal advocacy charities in conducting measurement and evaluation on their interventions. This organisation could also start conducting highly empirical cost-effectiveness analyses of various interventions and charities a la Givewell.
Despite this, I have some large uncertainties on the cost-effectiveness of the last animal dollar, the popularity of the Food System Research Fund amongst researchers, and how easy it is to influence funding.
There are approximately 70 billion land animals and up to 2.3 trillion wild fish killed for food each year. Most land animals are chickens, who experience high levels of suffering throughout their lives whether they’re an egg-laying hen or a broiler used for meat. Some estimates believe that 74% of all farmed land animals are in factory farms, with this number being even higher for farmed fish. In addition, the number of animals killed each year for food is growing, driven primarily by an increase in chicken numbers and increasing wealth across the world.
The farmed animal welfare (FAW) movement is trying to reduce the suffering of farmed animals, and support interventions that either improve the welfare of farmed animals or reduce the number of animals being warmed. Estimates by Farmed Animal Funders (FAF) puts the farmed animal movement at a size of approximately $200 million per year of funding, which has been steadily growing for the past few years. Despite $160 million of this being spent in the UK, Western Europe and the US, meat consumption for these three countries and regions still seems to be stable or still rising. However, there has been considerable success in corporate welfare reforms, with over 2,400 commitments since 2010.
This juxtaposition in relative success in improving welfare but failing to affect total meat and dairy consumption potentially points to some particular difficulties in improving farmed animal welfare. Specifically, we have very little evidence on most animal interventions, specifically around ones that affect consumer preferences or non-corporate welfare institutional asks, e.g. the impact of media, different messaging strategies, humane education or grassroots movement building. On the other hand, we have reasonably good evidence that corporate campaigns for chickens lead to significant improvements in welfare.
This intervention research seems extremely important to conduct, as ideally we want an animal advocacy movement that pursues multiple theories of change and is robust to specific failure modes of one strategy. In addition to novel research, measurement and evaluation of our existing work can also be extremely informative in future funding allocation.
Research now also has the benefit of providing early value of information, which will assist in resource allocation. Rather than do intervention research in the future, doing it now has the benefit of counterfactually influencing the funds that might have been misallocated up until we conduct future research.
How cost-effective could this be? A very rough BOTEC
It’s quite challenging to estimate the value of research, but a plausible yet rough BOTEC might look like:
- Conduct research that affects the total allocation of farmed animal funds by 2%.
- I would estimate that of the current total of $200m per year spent on FAW, about $60 million comes from “effective animal advocacy” or the EA-aligned FAW movement. This is based on $40 million per year coming from Open Phil, $10 million from ACE and another rough $10 million from EA Funds and FAF (but I have no idea about this last number really). I think this amount will be easier to influence via research, so it might be 5% ability to influence $60 million which is still roughly $3 million.
- However, the total money spent on FAW is projected to grow - possibly up to $260 million in five years time.
- This is based on a very rough extrapolation from the Farmed Animal Funders report that predicts a $20 million increase from 2020 to 2021 (see figure above). I assume this wouldn’t continue year on year, but I could be wrong.
- It also seems in line with hiring for the Open Phil FAW team. Their 4-person team distributes approximately $40 million per year (based on naive calculations from their grants database) and they are roughly doubling in team size. Therefore one should naively assume a rough doubling of funds, which would lead to disbursing $80 million, an additional $40 million over the next few years.
- Open Phil said they want to triple outgoing funding by 2025. This is across all of Open Phil though, rather than just FAW, so keeping our previous assumption of doubling might be reasonable.
- Potential future funds influenced = 2% of $230 million (average of $200 and $260 million) x 5 years = $23 million
- Cost of this research = $2 million over 5 years to find something good enough to change resource allocation by 2%. I’m highly uncertain about this factor, and I think it could very plausibly be as low as $200,000, but also as high as $5 million.
- Leverage factor (or rates of return) = $23million / $2 million ~ 12x leverage.
Average FAW spending per year, over next 5 years
Extrapolation from current trends
Percentage of spending that R&D could influence
Average spending that could be influenced per year
Total spending we can affect over 5 years
Total Cost (USD) over 5 years
Leverage Factor (benefit/cost)
To put this in terms of animals helped, we can consider the marginal improvement in resources spent.
- As a toy example, we can very roughly estimate that 2% of the funding influenced was being spent on interventions that were 10% as effective as corporate cage-free campaigns, before the research took place.
- With improved research, we can model this 2% (or $23 million) being moved to interventions that are 30% as effective as corporate campaigns, which is another factor I’m quite unsure about (and will largely determine how cost-effective this work is).
- This means we’ve generated a 20% improvement in funding allocation x 41 chicken-years affected per dollar, leading to an (extremely rough) BOTEC of 8.2 chicken-years affected per dollar spent.
- I’m very uncertain about how cost-effective the last animal dollar is, so I think this 20% value could very reasonably be anywhere between 5-50%. This would significantly determine how cost-effective this intervention is overall, so grantmakers who know more than me can judge how accurate these numbers might be.
How much do we spend on R&D at the minute?
FAF predicts in their state of the movement report that approximately $6.5 million of the total FAW $200 million is spent on either research or measurement and evaluation (see chart below). This is only 3% of our total spending per year, which seems far too low.
In the FAF report, it also states that organisations have relatively limited focus on monitoring and evaluation, around 1% of our total spending, which also seems lower than ideal.
However, in this comment, Kieran Greig from FAF estimates that animal advocates spend about 5-10% of our total annual budget of approx. $200 million on research, which is a bit higher than the 3% estimated above. Whilst 10% of our spending on research might be an adequate amount, 3% seems too low. Kieran also notes that given R&D spending on global health, approximately $3.5 billion, the animal welfare evidence base is growing <1% as fast as the global health evidence base.
He also adds that “Furthermore, global health has been around much longer, so plausibly the difference in sizes of the respective evidence bases could be on the order of a thousand times.”
The research spending for animal advocacy might also be slightly misleading, as a lot of this is for farm animal welfare science, rather than intervention-focused research. In my opinion, intervention-focused research and R&D is even more neglected than animal welfare research.
Who else is working on this?
The Food System Research Fund (FSRF) is working on this, with a fairly relevant request for proposal on addressing knowledge gaps related to existing advocacy techniques. Specifically, they are seeking to solicit ““proposals to examine how actual purchasing behaviour by consumers and food service companies is influenced by commonly used advocacy techniques. “ However, some potential limitations of this request for proposals from FSRF include:
- It seemingly only solicits research on existing or commonly used advocacy techniques, which means they miss out on innovation into new advocacy methods or less frequently used methods. This seems potentially quite important, as the best ways to help animals might still be undiscovered.
- Another drawback is the sole focus on “actual purchasing behaviour by consumers and food service companies”. Maybe this is a nitpick, but there seems to be no mention of interventions that could influence government policy or procurement. There is an example given about examining the impact of government policy on the reduction of sales, but seemingly the “link” before this is missing. This seems quite important, as it’s likely that interventions that influence government policy or regulations have the greatest possibility to create large-scale change for animal welfare.
- The FAQs list that they generally provide grants from several thousands to $30,000, although they will theoretically consider any size grant. I think it’s likely that population-level RCTs could cost much more than this, and it’s unclear if FSRF is open to this, which might dissuade potential researchers.
That said, they’ve funded some great research on this vein, e.g. work on the effectiveness of documentaries and an RCT on animal advocacy messaging. However, it’s likely that the funding being distributed by FSRF is still small relative to total farmed animal welfare funding, so an additional funder in the space could still add substantial value, particularly with more expensive population-level research projects or neglected advocacy methods.
In addition to RSRF, the EA Animal Welfare Fund has funded a small amount (only two research projects in the last round) of intervention research. Finally, Open Philanthropy’s Farm Animal Welfare team is hiring (or hired?) a Science Program Associate to assist with science-related grantmaking. However, in the job description, the mentions of science focus solely on animal welfare science (e.g. rates of layer hen keel bone fractures) or research into animal product alternatives. This suggests to me that animal advocacy intervention research is something that the FAW team is not considering funding to a large degree, which seems like a slight oversight.
Increasing the amount of funding towards animal advocacy R&D might be trivial (e.g. a funder allocated $Xm towards this area), but actually increasing the amount of research might be more challenging. For example, it’s not obvious that there are enough willing and able scientists or researchers who want to take on these animal advocacy research projects, so additional funding might have little marginal value. If this is the case, then building the pipeline of potential researchers to tackle these problems would be a high-leverage starting point. For example, Open Philanthropy already funds the Reducetarian Fellowship to support undergraduate students in the US working on ending the consumption of animal products.
However, this could be expanded to focus primarily on research e.g:
- A fellowship to provide Master’s or PhD funding support for talented individuals focused on animal advocacy research, such as agricultural economics or psychology (similar to Open Phil’s AI Fellowships).
- Additionally, funding summer research fellowships, like the the Stanford Existential Risks Initiative, might prove a strong way for a large number of early-career individuals to test their fit for research and possibly enter research departments at universities or animal advocacy organisations.
Besides building the pipeline of potential researchers, there are several low-hanging (and several more ambitious) fruit that animal-focused funders could do to increase the amount of R&D being done on animal advocacy interventions:
- Requests for proposals to highlight interest in the area, similar to that of FSRF.
- BERI for animals - an organisation that supports existing university research groups by providing support with things that are hard to do within bureaucratic university contacts e.g. hiring, purchasing and admin support.
- Providing restricted funds for animal advocacy organisations to hire specific measurement & evaluation staff, as most global health and development organisations have.
- Funding a measurement & evaluation initiative, such that existing animal advocacy organisations can be awarded unique grants for M&E related work.
- Incubating a new charity to support animal advocacy charities in conducting measurement and evaluation on their interventions. They could also start conducting highly empirical cost-effectiveness analyses of various interventions and charities a la Givewell.
- Incubating a new animal advocacy R&D university group, to attract talented students from a range of subjects e.g. psychology, economics, animal welfare science and political science.
Some specific ideas for research projects, copied from this megaprojects post, include:
- What impact does popular media coverage of undercover investigations have on animal product consumption?
- Vegan outreach - Similar to Mathur et al.'s (2021) work on documentaries
- Long-term impact of consumer-focused advocacy / humane education on animal product consumption
- Testing various forms of mass marketing of alternative proteins in populous low and middle-income countries
- Build on Sentience Institute's recent experimental work on the complacency vs momentum debate, which is a crucial consideration that seems very underexplored yet is very tractable given the resources. The results would thus inform what types of ambitious projects are most impactful. An example of this can be seen by Harris, Ladak and Mathur (2022)
- Protest impact on consumption (e.g. doing protests outside supermarkets and measuring any changes in purchasing, randomised across a state/area)
- Understand price elasticities of demand of conventional meat for countries with highest meat consumption (which might determine how important welfare reforms to raise the price of products are - some comments here)
- Build on some this work by READI e.g. behaviour change RCTs
Sources of uncertainty
Nothing in this piece is novel, and I would be surprised if major EA animal funders hadn’t considered most of these options already. Therefore, it’s very possible I’m missing something crucial that means making this R&D happen is harder than I think (e.g. the pipeline of talented researchers just isn’t there).
Otherwise, some major sources of uncertainty I have are:
- How much research you would (on average) have to do to have a 2% chance of shifting total funding towards another intervention. This is similar to my uncertainty below, but with additional open questions around how hits-based this research is likely to be. For example, it’s very plausible you commission $5 million of research without finding any major breakthroughs, but likewise you could also find it within the first $500,000. Therefore, I have quite wide confidence intervals on the expected value of advocacy R&D.
- How easily affectable the total amount of animal advocacy funding (approx. $200 million) is for new research. I estimated that 2% of total funding could be influenced by $2 million of new research, but in reality I have no idea how likely it is. Given that $60 million of animal funding is from EA sources, this seems a priori plausible as EA funders should hopefully update towards more effective interventions. However, it’s also a known flaw of CEAs to think that 1% (or thereabouts) is an achievable figure, when the reality might be an order of magnitude less.
- How cost-effectively the “last dollar” is spent for animal advocacy, as this will determine the overall cost-effectiveness of this work.
- How the Food System Research Fund is going e.g. are they receiving almost no worthwhile proposals or are they up to their neck with good research. This will largely determine how neglected I think this problem is, although I still believe they could expand their scope to cover more research.
- How big the pool of researchers who are both willing and capable enough to take on some of these researchers is. This will largely determine the interventions to prioritise e.g. should we build capacity first, or do we start funding projects straight away.
Saying all this, it seems like total global health spending is $8.3 trillion, so a percentage of total spending, R&D only makes up 0.1%, which seems very small but there’s potentially diminishing returns to research once you’re spending so much.
Thanks for sharing this submission! I just wanted to chime in with an idea I sometimes think about. I sometimes notice that a lot of the research on animal advocacy is targeted at grant makers who need to decide which projects to fund. But people who execute these projects also need to make a lot of strategic decisions and I suspect that the research that can support these decisions is more neglected. I reckon this might be because 1. research is costly 2. research has a lot of positive externalities 3. animal advocacy organisations are much more numerous than grant makers so it's more difficult to coordinate to provide this kind of "public goods".
Because of these reasons animal advocacy organisations rely on their own experiences to learn lessons. There are limits to how much you can learn from this since the sample size is very small. It's hard to predict what kind of work tests would best predict the performance of a campaigner when you have hired only 10 campaigners at most. Or it's hard to understand what kind of campaign strategies are better when you have run 50 campaigns. So I sometimes wonder whether it would be better to have well-funded researchers that reach out to animal advocacy orgs, learn about their research needs and help them.
Though a major problem with such a suggestion would be ensuring accountability. It's hard to ensure that this kind of research is really useful when animal advocacy organisations are not paying for it.
[doxxing myself - I wrote the submission above!]
Agreed Emre, I think that's a good point and probably something I didn't consider much for the purpose of this piece. I guess one factor that makes it less valuable for organisations conducting this research internally is that they've got pretty strong incentives to find good results, otherwise it means their work isn't effective. I think this can be a reason for people to measure the wrong thing, or otherwise not conduct research that really gets at the heart of whether they're being effective or not.
Also agreed about not being able to extrapolate much when there is small sample sizes. Something I've found quite hard for campaigning-related research is the fact that it's so context dependent that it's unclear how much you can extrapolate from even 1000 campaigns if they've been conducted in 20 different countries working on 3-4 different issues. Obviously somethings will generalise, but some won't! And it's hard to tell which one is in which camp a priori.
And your last point is another tough thing about research. It's well and good conducting great research - but often it can slip under the radar or people don't actually implement the recommendations. I think this is a time when external bodies (e.g. charity evaluators or funders) can be good at holding organisations to account on doing the most impactful things.
As an AI Safety-exclusive person, I'm definitely really glad to see more emphasis being put on the Advocacy part of Animal Advocacy. Advocacy R&D is definitely undersupplied skill pools worth funding.
I very much want to live in a world where an animal advocacy person can take one look at an AI safety person and tell them everything that they're doing wrong, after talking for only 5 minutes.