The AI boom could see anywhere from $37 billion to $100 billion in annual philanthropic spending over the next few years. The question is, how can this money be spent effectively? A new blog series from CGD makes the case for highly cost-effective areas that philanthropists may want to consider funding. This fifth blog in the series makes the case for funding education.
If I worked at a frontier AI lab and just became an overnight billionaire, my first thought would be to do something simple and give it all to GiveDirectly or GiveWell. But organisations in the business of delivering real-world, non-digital services are constrained in their growth by the difficulty of scaling organisations. Even GiveDirectly, which delivers one of the simplest things imaginable, just cash, think they could only double their spending in a year. With some hard work they might be able to scale to 10x, but that still couldn’t absorb most of the potential wave of philanthropy.
So where next? I work in education, so that might be my next move. Why education? I have some personal motivation: my parent was a teacher. GiveWell and Coefficient Giving have investigated education as a cause area in the past, and seen some signs of promise, but were left with too much uncertainty to recommend major grants over safer bets. I’ve always seen the case as being a stack of the labour market earnings returns (to both school and learning), and the social benefits, which are both hard to value. Welfare analyses in the US put interventions targeting children systematically above those targeting adults. School-building programs in middle-income countries, far from the most cost-effective interventions, have marginal value of public fund ratios (similar to a benefit-cost ratio) of over 10 to 1, and efforts to improve the quality of schooling can do similarly. These future benefits might be even more uncertain to an AI billionaire with short timelines who thinks we’re heading to a very weird future. Perhaps mass human capital will soon be economically irrelevant. Perhaps we’re back to an even more vague justification based on self-actualisation. Fundamentally, I think that regardless of what the future holds, it is now globally normative that we should offer all children the opportunity to discover things about the world and about themselves, to access the knowledge and skills that they need to engage in the world, and work out what talents they have and how they can make a contribution to others.
How to do that concretely? I might start by scattering some no-strings money across some of my favourite NGOs, MacKenzie Scott-style (albeit with the important caveat that all of these organisations work exclusively in low- and middle-income countries). So here’s my list: 34 organisations doing important work, with an eye for scale and cost-effectiveness. This is not a comprehensive list. It’s some of the organisations I have some passing familiarity with, who I’d be willing to place a bet on. I have no doubt left out many other excellent organisations. Do let me know! My rough estimate is that their total combined annual spend is about $650 million, so perhaps you could realistically and feasibly aim to double the overall size of their combined operations with annual spending of that order.
What do you do after that? The real route to scale and sustainability in education is through governments. It might not be the single most cost-effective way to spend money, but you could spend an almost endless amount in bringing all the primary and secondary schools around the world up to a basic standard, with clean and safe buildings, toilets, and books. One concrete way for an individual to engage directly in funding a government could be through the Foundation for the Global Partnership for Education. Alternatively, a high-leverage move is investing in research to inform government spending, such as through CGD (full disclosure, I work here).
Here’s the list of organisations (mostly nonprofit NGOs and some for-profit social enterprise companies).
Foundational learning
- Pratham: Maybe the best education NGO in the world? Pratham pioneered citizen-led learning assessments that have been adopted in 15 countries. Pratham also developed the "Teaching at the Right Level" (TaRL) approach that is one of the "smart buys" in education, which has also spread to over 30 countries. The approach has been tested through multiple partnerships with J-PAL and Nobel laureates Duflo and Banerjee.
- PAL Network: The umbrella organisation supporting member organizations to conduct those citizen-led learning assessments. Policymakers underestimate the scale of the learning crisis, and these independent learning assessments help bring it into the spotlight.
- TaRL Africa: A joint venture by Pratham and J-PAL, purpose-built to scale Pratham’s TaRL approach across Africa (currently in 16 countries, mostly with national governments).
- VVOB: A Belgian organisation focused on working with ministries of education, through government systems, and with teachers and school leaders, including importantly acting as a key partner to TaRL Africa in embedding that approach in several countries. They have been admirably open to collaborating with researchers and publishing impact results even when they are disappointing.
- Youth Impact: An innovation engine working on adapting promising interventions such as TaRL through different approaches such as one-to-one phone tutoring, and committed to testing with a stream of academic publications, and to my mind, great credibility from pivoting away from an earlier intervention that was evidence-informed and promising but didn’t work out.
- Alokit: An Indian NGO supporting school leaders to improve their coaching of teachers, with a major focus on foundational learning and improved pedagogies such as TaRL (disclaimer: I have a paper close to finished studying their model).
- Global School Leaders: The international umbrella organisation supporting national principal and headteacher training organisations in 17 countries.
- Imagine Worldwide: Scaling an adaptive learning software nationwide in Malawi. The software has won awards and is supported by several individually randomized RCTs. I’m looking forward to the larger multi-school RCT.
- Luminos Fund: Runs effective catch-up programmes for drop-outs in the poorest places in the world. A recent RCT showed large effects.
- Room to Read: They publish books in local languages, build libraries, and train teachers in evidence-based phonics instruction. As someone obsessed with books as a kid (and now whose own kids are hard to drag away from their latest library find), I find it hard to comprehend what it must be like to be a kid in a rural village where there are no story books. None in the school library, which only stocks a few textbooks, and no public library at all.
Schools and access
- BRAC: Not just an education NGO, but nonetheless the largest NGO in the world, and the largest non-state provider of education in the world, by some distance.
- The Citizen’s Foundation: One of the largest school networks in the world, running over 2,000 schools in Pakistan, with an all-female teaching faculty.
- PEAS: Build and run quality secondary schools in underserved rural African communities. I studied the network as part of my PhD, finding them to be extremely well-managed.
- Street Child: One of the best performers in Liberia’s Partnership Schools pilot evaluated by researchers at CGD.
- Rising Academies: The other top performer in the Liberia Partnership Schools pilot, who continue to pursue evidence-based structured pedagogy, and innovate with AI, and with COVID distance learning (also evaluated by CGD researchers).
- World Bicycle Relief: Have distributed nearly a million bicycles. I love bicycles on principle, but there is also very good evidence that giving them out is a cost-effective way of improving access to schools in rural settings with few schools and little alternative transport.
- Friends of African Village Libraries. A relatively small group, but I basically think it’s impossible to spend too much money on village libraries in low-income countries.
Early childhood
- Lively Minds: A volunteer-led preschool programme with rigorous evidence of impact from Ghana.
- Rocket Learning: WhatsApp support groups for parents and staff at preschools in India, with evidence finding impacts at very low costs of about $1 per pupil.
Health and nutrition
- Evidence Action / Deworm the World: Deworming is still one of the cheapest levers for school attendance that exists.
- Food 4 Education: a Kenyan NGO feeding half a million children daily, at a cost of $0.30 per kid. Meals are not the most cost-effective way of improving learning, but there can be no deadweight loss in feeding children. And meals get students to school.
- Mary’s Meals: a global school-feeding charity based in Scotland, feeding around three million children a day across 16 countries.
Lead exposure
- Pure Earth: You might be surprised to see lead exposure listed as an issue here, but it matters hugely for education. Half of children in developing countries have some exposure, which overall accounts for huge negative impacts on learning according to CGD research. Pure Earth is the largest and longest-running global organisation focused on tackling the issue. You can watch CEO Drew McCartor’s TED Talk here.
- Lead Exposure Elimination Project: Half the countries in the world have no regulation on lead paint. LEEP has already changed that in multiple countries in just five years of operation, and are rapidly scaling their work across dozens of countries. Their estimated cost per disability-adjusted life year averted is absurdly low.
- LABRI: A brand new organisation focused on tackling unsafe lead-acid battery recycling, which is responsible for perhaps a third of the burden according to CGD research.
- Pb Action: Another brand new group focused on the same question—unsafe lead-acid battery recycling. Given the size of the problem, there is no risk yet of this space getting overcrowded.
- LeRA. Yet another brand-new organisation focused on research on lead exposure but with a broader remit.
Educational media and technology
- Sesame Workshop: I loved Sesame Street as a kid, so I love that we have evidence that it improves educational outcomes. Mass media has inherently high cost-effectiveness if you get the content right because the marginal cost of delivery is zero. Sesame Workshop does the local adaptation.
- Ubongo: Is the homegrown African version of Sesame Street, reaching millions of children across the continent.
- Mindspark by Educational Initiatives: The stand-out education technology project, with big effect sizes found in the lab and the scale-up in India.
- Taleemabad: At the Skoll World Forum this year, everyone was talking about AI, and whenever I asked anyone about a good concrete example of an organisation using AI well for education, the answer was Taleemabad. We have a lot to learn in this space, and Taleemabad seems to be leading.
Child safety
- Raising Voices: There is an epidemic of violence against children around the world, and we have very few RCTs of interventions designed to reduce violence in and around schools. Raising Voices developed the programme tested in one of the first, the “Good School Toolkit.” Their CEO Dipak Naker spoke recently with Gabriela Smarrelli and David Evans on the CGD podcast.
- Coalition for Good Schools: The global practitioner network (incubated by Raising Voices) convening the evidence on preventing violence in and around schools, and advocating for greater action.
Skills
- Educate!: Improving the work readiness of high school students in East Africa, Educate! is rare in both running multiple RCTs on their model, and working directly with and through government systems.
