That's amazing; thank you for this initiative and fundraising for the EA AWF! 

Why did you assume hiring the best candidate would save senior staff 66 h/year.

The number, is the difference between the first and second candidate in $ cost per h saved (given their salary and how much time they save). The difference would be $66 per h. Later, I accidentally omitted the $ sign in the text, and that indeed created a mistake in further calculations. It turns out that making BOTEC late in the evening is not a good idea, in my case. :) Thank you for catching that error!

To refine the calculations by fixing the error you spotted and adding more considerations:

I say earlier. 
“Let's assume the 2nd best candidate is half as good as the 1st and therefore saves us half as much time. Instead of saving us 35h per week (40h - 5h management, meetings, review etc.), they save us 17.5h (requiring much more management and oversight to get the same results, which I actually think is conservative), for the same up to $120k spent on salary and benefits.

In the first case, we get 35 × 52 workweeks in a year = 1,820h, and in the second 910h. The cost-benefit analysis is $65 per hour for the first candidate and $131 per hour for the second, with a difference of $66 per hour.”

I was aiming to calculate the difference between the first and second candidates. The first would save 1820h (35h × 52 workweeks in a year), the second 910h (35h x 52 workweeks in a year). 
The cost-benefit ratio of that time saved is: for the first, $65 per h ($120000/1820 hours saved per year), and for the second, it is $131 per h ($120000/910h saved), so the difference is $66 per h. 
1820-910=910h difference in a year
And each hour, for the 2nd candidate, cost us $66 more than for the 1st.
910h difference, at $66 per h, the difference in cost is $60060. 
Indeed, in that time, a senior fund manager could in theory, create 60 active grantmaking opportunities (910h/15h) at a cost of $60060. 
So $60060/60 = $1001 difference in cost between candidates for generating one opportunity. 

But you are right; we cannot generate 60 active grantmaking opportunities in a year, no matter the time spent. If we had an unlimited time (something I didn’t assume in the RFMF estimate), I think we could generate more than $2M estimated in the RFMF post, but indeed not 60 opportunities. My guess would be somewhere around 20-30. If we take those numbers and follow your reasoning, there is room for saving from 300h (=20*15) to 450h (=30*15), which could be achieved by hiring a candidate at least 16.4% (=300/(35*52)) to 24.7% (=450/(35*52)) as good as the best. While we have to remember to take into account the higher cost of generating that opportunity in the case of the 2ns candidate. 

A significant limitation to that estimate is that we assume no increase in time when generating the next marginal opportunity. In fact, I expect that each marginal opportunity we generate will require a higher time investment to generate, simply because it will be harder to come up with ideas, find the right people who are not already busy, etc. So let me introduce this refinement to our estimate. For example, the first 10 opportunities may take 15h per idea, the next 10 can take 25h, and probably the next 10 would take significantly more, like 40h. If we take those numbers and the range for the number of potencial opportunities (20-30), the total time to generate 20 opportunities would be 400h (=(10*15)+(10*25)), and the time to generate 30 opportunities would be 800h (=(10*15h)+(10*25h)+(10*40h)). So the potencial of saving 400h-800h is generated by hiring a candidate at least 21.9% (=400/(35*52)) to 42.9% (=800/(35*52) as good as the best. While remembering that generating those opportunities by the 2nd best candidates would have a worse cost-benefit ratio. We also have to remember that in this case, we may need more than $2M for active grantmaking, which further complicates calculating the "better candidate to more funding trade-off". 

However, I will stop this estimate now, because the time for the AMA is running out and I have to get ready for the beginning of the holiday that starts in Poland today. :) If you have any comments about the calculation above, let me know. If I happen to have some free time after the holiday, I may swing back to finish and further improve the estimate, but I cannot commit to that, especially if it would trade off against vetting and selecting the best candidates in the current hiring round ;) Thanks for this exchange and all your questions, Vasco! 

Hi Vasco! I answered your question in the recent Ask Us Anything, and here, I have written a more extensive list of when I think it is better to donate to an individual charity and when to a fund. 

Hi Michael! You can find the announcement post and discussion in the comments here. 

Thanks for the question. This is not a question the fund has to consider very often - we're typically evaluating grants that would affect animals living lives we expect are negative lives. 

It's possible there are some cases where we're evaluating some interventions to reduce the number of farmed animals (e.g., meat reduction or farm prohibitions) where some of the animals who would not come into existence because of the reform would have otherwise lived net positive lives (some have estimated under particular ethical assumptions that cows raised for beef could be living net positive lives), but the vast majority of the impact would still be aiming to affect animals that are experiencing net negative lives and don't have a viable path on the table to achieve net positive lives instead.

To go a bit more in-depth and offer a more personal take rather than speaking for the AWF. Personally, even if I put on my 100% utilitarian hat, I would still have some uncertainty. First, I would need to have high confidence in:

1. Their lives being in fact, net positive. I think there is too much uncertainty about the welfare of animals and how to approach it depending on ethical theory. For example, if we give significant credence to negative utilitarianism or simply place a high weight on excruciating pain, then even if animals lead decent lives, slaughter at the end of their lives may outweigh it.
2. Methods to measure and aggregate their welfare, such as temporal welfare aggregation challenges.
3. Ability to determine where this threshold is and at which point life becomes net positive. Some claim that this concept is not even valid.

But yeah, if I was confident in all of that, or was risk permissive, with a 100% specific flavor of utilitarianism hat on, maybe I would.

But personally, I'm not sure I'm 100% utilitarian, and I have a more complex parliament, where some members/ethical theories say that it wouldn't be ethical (e.g., rights-based). I could imagine a case where more members would agree if, for example, there was no slaughter before natural death would occur for a given individual, and animals would die being completely anesthetized. Additionally, farms would be completely open, where an animal could choose to leave the environment they are in (where they are taken care of, but their products are taken away from them) and choose another one (where perhaps they are not taken care of but are free to fully express their natural behavior (e.g., where their offspring would hatch from eggs instead of eggs being taken away from them)). There are still dilemmas, such as whether truly informed consent is possible for animals, whether the choice to stay implies positive welfare or just status quo bias, and whether providing choice is sufficient for moral permissibility, etc. 

Then there is also the issue of whether we are obligated to bring into life beings who will lead net positive lives. I certainly don't act in accordance with that now, and I think population ethics is something I cannot solve, so I don't know what I would do. ¯\(ツ)/¯ Jokingly, maybe spend my donation budget to fund Peter Singer to figure it out :P 

One way to BOTEC this is by looking at how much time it saves us and what can be done with that time. Let's assume the 2nd best candidate is half as good as the 1st and therefore saves us half as much time. Instead of saving us 35h per week (40h - 5h management, meetings, review etc.), they save us 17.5h (requiring much more management and oversight to get the same results, which I actually think is conservative), for the same up to $120k spent on salary and benefits.

In the first case, we get 35 × 52 workweeks in a year = 1,820h, and in the second 910h. The cost-benefit analysis is $65 per hour for the first candidate and $131 per hour for the second, with a difference of $66 per hour.

As discussed in our room for more funding post, we currently believe we could conduct more active grantmaking to a value of $2M. If we assume that with 15h, a more senior staff member whose time we saved by hiring can generate an active grantmaking opportunity that costs $200k, and assuming its cost-effectiveness is $1.4 per DALY ( I took the RP CCM DALY estimates (which you helpfully listed here in DALY/k$, and I reversed to be $/DALY), where the Cage-free Chicken Campaign was $1.4 per DALY.), that means 142k DALYs difference. In the 66h difference between candidates, we get 4.4 such opportunities, so 624k DALYs are lost due to hiring the 2nd best candidate. At $1.4 per DALY, that's ~$873k.

Therefore, if the 2nd best candidate is half as good as the first, we would need $873k more to offset it. This could be a conservative estimate, because a half-as-good staff member might simply not generate as good evaluations no matter how much management time they get. Or it could be liberal because we may need more than 15h to generate the next marginal active grantmaking opportunity, or the 2nd best candidate could be more than half as good as the first. I think a range of $500k-$800k is reasonable.

That was a fun exercise; thanks for your question! 

We are excited about efforts to increase the amount of funding that goes to high-impact animal interventions. That being said, we believe there are as many, if not more, promising opportunities to increase funds from these other sources, such as:
a) less effective animal sources, supporting work of animal-focused effective giving and fundraising initiatives such as FarmKind, or Farmed Animal Funders, and cross-cause ones, e.g., Effektiv Spenden and others. I believe AIM had a report offering an impact evaluation of those, but I cannot find it now. 
b) less effective human sources, such as leveraging government R&D funding to be redirected to alt protein. This had significant successes, as described by Lewis in his new newsletter “6. Putting Green into Going Green. Governments invested over $200 million into research and infrastructure advancing alternative proteins, including in the US ($71M via DOC, DOD, and Massachusetts), Denmark (DKK 420M / $59M), Japan (¥7.87B / $51M), the UK (£27M / $34M via two grants), the EU (€12M / $13M) and Beijing (80M Yuan / $11M). New alternative protein research centers, funded by the Bezos Earth Fund, opened in London, North Carolina, and Singapore.” We also think that influencing climate philanthropy has a lot of potential. 
We haven’t evaluated the two methods you described, and I’m not aware of any such estimates, so I cannot comment on their effectiveness. But I think that in any scenario, those interventions I mentioned would be better on the global net, species-agnostic welfare than, e.g., moving from the best interventions helping humans to the best ones helping animals. 

Thank you for this thoughtful question and for your kind words about the Animal Welfare Fund! You raise an important point. Let me break down our approach:

First, we don't operate with fixed portfolio allocations or minimum percentages per species. Instead, we aim to maximize the marginal impact of our grants based on our best current understanding. This means evaluating each opportunity on its own merits and seeing if it is above our bar. More about our bar here. 

Secondly, it is worth noting that purely theoretical calculations often differ significantly from practical funding opportunities. While back-of-the-envelope calculations might suggest allocating a large percentage to certain species (like shrimps), we simply don't see enough promising, implementation-ready opportunities in those areas to make such allocations feasible. Historically, we were more limited to the applications we received, but recently we started doing more active grantmaking to generate those opportunities in areas that are cost-effective but neglected, and in 2025 we plan to further invest in it. 

Even still, if those opportunities existed, I think it would be unwise to make decisions purely based on those naive utilitarian calculations. I say naive, referring to the difference between an actual cost-effectiveness and estimated cost-effectiveness. If I knew the actual cost-effectiveness of given interventions, that accounts for all uncertainties:

• Empirical - e.g., how many shrimps are we actually affecting?
• Moral - e.g., how to trade off different intensities of pain and what moral weight should be assigned to different species?
• Epistemological - e.g., how to account for far-future effects vs near-term effects of animal welfare interventions, given complex cluelessness?

that gives me a true number for cost-effectiveness, a “god comes from the sky” kind of situation, then I would rely on it. However, any estimate of cost-effectiveness is going to be a naive one and merely a very uncertain estimate that may miss those important uncertainties. Additionally, I would refer here to the timeless classic “Why we can’t take expected value estimates literally (even when they’re unbiased)” by GiveWell. While AWF's approach is different in some places than the one outlined in this GW blog post, I think the main point stands. They conclude that:
“I feel that any giving approach that relies only on estimated expected-value – and does not incorporate preferences for better-grounded estimates over shakier estimates – is flawed. Thus, when aiming to maximize expected positive impact, it is not advisable to make giving decisions based fully on explicit formulas. Proper Bayesian adjustments are important and are usually overly difficult to formalize.”
In light of that all, I think that we have imperfect information and too much fundamental uncertainty to justify extremely undiversified allocation, even if explicitly utilitarian calculation would point to that. 

Additionally, in practice, our fund managers bring diverse perspectives on for example how to weigh speculative versus evidence-backed approaches. This natural diversity helps ensure we maintain a balanced portfolio between proven interventions and those with high expected value but less certain ones.

Currently, we're working on refining our strategic framework, which may introduce additional allocation considerations. That said, our focus on neglected species and interventions already creates an implicit prioritization - we rarely fund work focused on cattle welfare, for instance, as other funders adequately cover this space.

For each grant we consider, we assess whether it meets our cost-effectiveness bar, which is influenced by other opportunities we see in our pipeline. This approach allows us to remain flexible and responsive to the most promising opportunities while maintaining high standards for expected impact.

I wish I had a more quantitative answer at this point. We have begun tracking grant impact, and forecasting its outcomes using a system that categorizes grants into four possibilities: successful as planned, successful pivot, unsuccessful due to theory of change, and unsuccessful due to execution. Once we've collected more data through this system, we'll be able to provide more precise numbers. :)

For now, I can say that modifications to original outcomes happen fairly often - my rough estimate is in about 30% of cases. These modifications can involve either scaling back to more modest goals or, in some cases, expanding to more ambitious ones.

Generally, we view pivots positively when grantees adapt their outcomes based on new information and the modified outcomes have led to (or are likely to lead to) meaningful impact. In these cases, we increase our confidence in the grantee's ability to execute this type of work while decreasing our confidence in the assumptions underlying the original theory of change. When grantees don't deliver their planned outcomes, we look for evidence that they've learned valuable lessons that will help them either develop more realistic expectations and plans, or improve their tactics to better achieve their goals. 

While we appreciate all the efforts that advocates undertake to improve the plights of animals, we do take track records into account when evaluating subsequent applications. Our tolerance for "misses" before deciding not to fund a grantee depends on several factors, such as our priors about the effectiveness of their work, or the ambition of their undertaking - two misses from a grantee that has been achieving significant impact for years but is now struggling with an ambitious campaign is different from a grantee who misses twice on their moderate goals and hasn't had any positive track record before.

Ultimately, how we balance accountability versus flexibility is highly case-dependent, taking into account the full context of the grantee's work and circumstances.

As AWF, we haven’t made any direct comparisons between AWF and GW, so we wouldn’t be able to answer that. While I read your cost-effectiveness analysis, and I’m grateful for the work you do to publicly estimate that for a range of charities, I won’t have the capacity to review it in depth in order to point out specific disagreements if they were to occur.