Hide table of contents

How do we get more EA charities started? There’s a good case that charity entrepreneurship is high impact for EAs, but it seems not many are starting them. Part of the reason is that it’s intimidatingly hard. Not only do you have to have multiple rare and difficult skills, but you also have to choose a good idea to begin with. And if people are put off by the uncertainty of career selection, that’s nothing compared to the sheer ambiguity of all the potential interventions one could run. That is why we are starting a new program called Charity Entrepreneurship. We will make yearly charity startup recommendations, much like what GiveWell does for donations and 80,000 Hours does for careers. This research will be accompanied by an incubation program, similar to Y Combinator, that will teach people the prerequisite skills to start a nonprofit. In this post I’ll explain more about the rationale, details, history, and plans moving forward. 



The first project undertaken by Charity Entrepreneurship (CE) was 6 months of research in 2016 by 4 full time staff (~24 staff months). This resulted in Charity Science Health being founded, which has since received two GiveWell incubation grants and assisted over 200,000 families in India. This was the primary goal of this project and we feel CE succeeded. However, in a another way, CE was much more successful than we anticipated. Another project, Fortify Health, was founded, helped by our intervention recommendation, mentorship, and seed grants. Not only that, but several other EAs and non-EA charity founders showed a large interest in founding high impact ideas, such as the ones we had researched, although we did not have the ability to support more than two groups at the time.


How it will work

The organization’s time will be broken down into three sections. The first section will be systematic empirical research into an intervention area focused on finding the highest impact gaps to found a charity in. This will result in a list of “top charities to found in X area”, much like our published list from our last round. The second part will be a two month summer incubation program equivalent to about a summer term university course load, focused on building the practical skills need to found a charity. The third section will be a seed grant and weekly mentoring sessions for the organizations that are founded out of the program, similar to the support we were able to give Fortify Health. We believe this program has potential to found 1-3 GiveWell incubation/ACE recommended equivalent charities a year. 

Each year a different area will be chosen to focus on, with research being aimed at prioritizing specific possible charities within that area, as opposed to actively comparing across them. This is for multiple reasons. Firstly, in the time it would take to feel very confident in prioritizing one cause area over another, we could have incubated a good charity in each of the top cause areas. Secondly, there is a limited pool of people interested in starting organizations in each area, so focusing on putting out marginal recommendations in one field will lead to less output than switching between them periodically. Thirdly, given the extremely uncertain nature of doing good, rotating between cause areas makes the impact more robust in case one of our assumptions or beliefs are very wrong. This perspective also aligns well with the recent writing on epistemic humility which we found persuasive. The area we are running the first year on is animal rights, and for the next year we are considering far future and mental health, among others. We already have recommendations listed for global poverty and count CSH and Fortify Health as a somewhat informal first year of running this project. 


Case for impact

The case for why charity entrepreneurship is effective has already been made here, here, here and here. I’ll make the case for how Charity Entrepreneurship as an organization has impact in this section. 

CE’s impact hinges on two factors - the counterfactuals of the staff running the organization and the counterfactuals of the participants. My personal counterfactuals would be starting another direct nonprofit myself. The benefits of CE compared to this is that, while I think I could start another good charity, it takes about 3-8 years to take a charity from starting to being able to run without the founders. On the other hand, with CE we would be able to launch an expected 2-3 charities through our program every year. While I may be an above average charity founder, I do not predict that I am 6 to 24 times better than the incubatees. 

Which brings us to their counterfactuals. Some people would undoubtedly start charities without our aid, but many wouldn’t. Fortify Health for instance, has said that they think they wouldn’t have started their organization without my mentorship and initial funding. Making the leap from structured jobs to the uncertainty of entrepreneurship without a guide and some initial money is intimidating enough to turn off a lot of people. Providing a structure, education, guidance, and seed funding can help a lot of people gain the confidence to make the switch. Additionally, the support makes it more likely that they’ll run a higher quality organization, making better decisions and having a broader skill set than simply jumping in. The assistance will also help them stick with it at the beginning when things are particularly challenging. It’s easier to keep going when there’s a setback if you have a coach encouraging you to keep going. Part of the inspiration for this is to be what we wish we had when we had started. 

Aside from the direct effects of causing more people to join, there’s the benefit of choosing a better idea. Most startups are started based on jumping on the first opportunity the founder sees rather than systematically comparing their options. This can sometimes lead to good ideas but is dominated by comparing between alternatives, which I go into more deeply here. Furthermore, far more time and expertise will be put into prioritizing the interventions. Even if the founder might have put some effort into comparing their best options, it’s hard for an individual to put in the multiple person-years that will go into our research. Lastly, the intervention prioritization will be done by people who have specialized in and have a track record in choosing good ones. All of this combined will lead to higher impact charities getting started than otherwise would have. 


How you can get involved

If you want to help this project you can:

  • Apply to our incubation program. The next session will be held in the summer of 2019. To find out when applications open, join our mailing list.
  • Apply for a research internship. The internship will work on shallow intervention research focused on finding areas for new charities to be founded. For more details, go here.
  • Donate to us. You can see our budget here. Any additional funds past what is listed will go to more and larger seed grants, which will allow new charities more breathing room and encourage more people to join our program. To donate please contact peterqc@charityscience.com 
  • Talk to us at EAG San Francisco. Both Katherine and myself will be attending, so if you are interested in the program, we’d love to talk to you there. 






More posts like this

Sorted by Click to highlight new comments since:

Very nice - I've had people ask me before how to make a charity more effective, and it's always been somewhat uncomfortable to have to say that EA focuses more attention on evaluating the existing effectiveness of charities than on trying to help charities to become more effective. But this is one step better than just helping existing charities to become more effective, this is creating effective charities from the ground up. Bravo.

Awesome! So pleased this is finally happening.

This is amazing, congrats on making this happen!

Fantastic idea. There are a couple of similar "for-good" incubators in the UK: Year Here (yearhere.org), which focuses on charities / social enterprises, and Zinc (zinc.vc), which focuses on more-than-profit businesses. Neither are explicitly EA, and I'm sure plenty of the organisations coming out of them are focusing on sub-optimal causes and/or interventions, but they may be interesting nonetheless. I know people at them if you want to talk to anyone.

Awesome! Do you also have plans to assist EA founders of for-profit social enterprises (like e.g. Wave)?

It is possible we will focus on this in one of our future years.

"We believe this program has potential to found 1-3 GiveWell incubation/ACE recommended equivalent charities a year." How many applicants are you expecting to mentor each year? Or put another way, what percentage of new charities you support do you expect to become top charities?

I expect ~10 people to attend the camp although I do not expect 100% of them will start charities (I would guess ~60% would). Out of charities founded I expect about 50% of them would be GiveWell incubation/ACE recommended. Although it would depend on the year and focus.

I expect ~10 people to attend the camp although I do not expect 100% of them will start charities (I would guess ~60% would)

So you mean you expect 6 different charities to start, or that 6 people will be involved in starting a charity, possibly the same one(s)?

6 people will be involved in starting a charity, possibly the same one(s)

This seems high compared to the startup equivalent. For example, I know at Entrepreneur First they take on 100 people and they form about 20 companies - don’t know how many get seed funded but 50% seems too high.

I would guess a lot of this depends on the number of people you take on. (e.g. if we took 20 people I do not expect we would get 2-6 effective charities.) I also would guess the odds of effective charities being founded if it was not picked from our pre-researched list would be much lower, something closer to 1/10 - 1/20.

Our estimates are mostly based on our experience with charities we have founded/supported in a pretty similar way to the above. I also am unsure how to generalize from for-profit to nonprofit space. I generally think the former is much more competitive.


Out of charities founded I expect about 50% of them would be GiveWell incubation/ACE recommended.

Is there anything more you can say about why you think this? (Feel free to ignore the question - I don't have anything substantial to say about why this seems high to me.)

Hey sorry for the slow response on this. I was waiting for some information to be published. I think my estimate would have been much lower before Charity Science Health and Fortify Health both becoming GiveWell incubated. Fortify Health in particular, I think is fairly representative of the program I plan on running, although the future program will likely provide more support than what I was able to give to their team.


No worries, thanks for the response. I've done a bit more digging around the outside view here and this no longer seems high to me.

Looks rational. Curious to know, what the program really offers. Economically, any activity is just a conversion of one type of assets into another. A charity is essentially converting money into -- e.g., a world without poverty, a world without animal suffering, etc., and in a sense, it's selling the world... and it is still a trade, just trading the delta of the distance to the goal...

More from Joey
Curated and popular this week
Relevant opportunities