Published: June 2023
In January 2023, GiveWell recommended an $855,000 grant from EA Funds' Global Health and Development Fund to the Agency Fund, which funds fellowships and projects that “expand people’s agency over their own lives.”1
This grant will fund the following activities:
- $630,000 will fund fellowships of $70,000 each to nine organizations that aim to improve livelihoods through "Youth Empowerment."2 Participating organizations will receive training in technology, data science, and social psychology, as well as GiveWell's approach to cost-effectiveness analyses and economic outcome measurement.3 At the end of the seven month fellowship, each organization will prepare a short writeup describing its program—including cost-effectiveness estimates and the evidence for those estimates—and requesting funding from GiveWell (if applicable).4
- $225,000 will support core operations of the Agency Fund.
We are recommending this grant because we think it may help us identify cost-effective funding opportunities in the livelihood improvement space. While we are interested in expanding the number of programs that we recommend funding which primarily improve income or livelihoods, we’ve typically had trouble finding livelihoods programs that meet our cost-effectiveness bar. We think the Agency Fund may help us identify cost-effective livelihoods programs by:
- Screening potential organizations to identify those which are more likely to meet GiveWell’s bar
- Helping organizations put together informative proposals for GiveWell by training them on how GiveWell evaluates evidence and models cost-effectiveness
- Saving GiveWell staff time in identifying potential organizations and sharing information about our approach to cost-effectiveness
- We also think it is possible that the training on technology, data science, and social psychology that fellows receive will have positive impacts on the cost-effectiveness of their organizations’ programming, although this was not a primary consideration in making this grant.
We are providing operating support to the Agency Fund because a) we think it may support the Agency Fund’s continued growth and ability to identify and support potentially cost-effective organizations in the future and b) to cover the costs of staff time associated with running the fellowship.
Our main reservations are:
- This is an experiment and we are unsure how likely it is that GiveWell will fund an organization participating in the fellowship.
- We feel uncertain whether either a larger or smaller investment in operating support for the Agency Fund might have been appropriate.
We think there is a 60% chance that we decide to fund at least one of the nine organizations participating in the fellowship for additional data gathering or scoping by December 2024.
GiveWell recommended this grant via our policy for small discretionary grantmaking. As a small discretionary grant, this funding opportunity did not receive the same scrutiny as larger grants we recommend. Instead, we more minimally evaluated the case for the grant and any potential risks or downsides.
|Agency Fund, "Our Approach"
Richard Sedlmayr, the Agency Fund, email to GiveWell, January 9, 2023 (unpublished).
“We generally want to bring fellowship cohorts to a shared level of knowledge in terms of data science, impact evaluation, technology, evidence, and social psychology. This typically involves some training but also quite a bit of collegial workshopping. In this specific fellowship, we would add modules on economic outcome measurement and cost-effectiveness, with a deep dive on Givewell's metrics.” Richard Sedlmayr, the Agency Fund, email to GiveWell, January 9, 2023 (unpublished).
- “Overall I propose as output documents one five-pager prepared by each org describing their own potential in givewell metrics / language, plus an independent agency fund report / summary / deck that synthesizes all we learned on our side about these grantees; about how they related to GW hurdles; and what might be done going forward.” Richard Sedlmayr, the Agency Fund, email to GiveWell, January 9, 2023 (unpublished).
- We expect to receive outputs from fellows in November 2023.