This is our current answer to the site’s one question. It ranks cause areas, not individual charities, by what the published evidence says a marginal dollar buys. Within each area we name the programs the specialist evaluators fund, because that is where the numbers come from.
Two rules for reading it. First, differences of 2x are noise; differences of 10x are real. Second, every figure below is an estimate made by someone else, with a date. We link to each one.
The ranking
| Tier | Cause area | Best-evidenced cost per outcome | Evidence quality | Room for more money |
|---|---|---|---|---|
| 1 | Child health in low-income countries | $3,500 to $5,500 per life saved (GiveWell, 2022 to 2024 average) | Strong: randomized trials plus large-scale program data | Large, and larger after 2025 aid cuts |
| 1 | Lead exposure | About $4.49 per healthy year of life (LEEP model, 2024); $1.66 per child protected (Founders Pledge) | Moderate: model-based, forward-looking | Growing: paint gap for 2026 filled, expansion open |
| 2 | Mental health in low-income countries | $21 to $25 per WELLBY (Happier Lives Institute); GiveWell estimates lower | Moderate: trials exist, effect size and duration disputed | Yes: StrongMinds projected a $6.9M gap for 2026 |
| 2 | Farmed animal welfare | Pennies per animal-year for corporate campaigns (Rethink Priorities, ACE) | Moderate: outcomes countable, follow-through uncertain | Yes: ACE’s fund is small relative to the field |
| 2 | Climate policy advocacy | Roughly $0.10 to $1 per ton CO2e in expectation (Founders Pledge on CATF) | Weak to moderate: counterfactual unknowable, politically exposed | Yes, but crowding at the best-known groups |
| 3 | Unconditional cash transfers | Near 1x the benchmark by definition; strongest certainty of any program | Very strong | Effectively unlimited |
| 4 | Domestic causes in rich countries | Rarely measured; cost per outcome typically 10x to 100x the above | Varies widely | Varies |
Tier 1: what your dollar can most reliably do
Child health. GiveWell’s four top charities, vitamin A supplementation (Helen Keller Intl), seasonal malaria medicine (Malaria Consortium), vaccine incentives (New Incentives) and bed nets (Against Malaria Foundation), are the most studied interventions in all of philanthropy. Their cost per life saved comes from trials replicated across countries, updated with current disease burden and program costs. GiveWell’s own bar for funding, as of May 2026, is 6 times its benchmark of doubling a very poor person’s consumption for a year; these programs clear it. After the 2025 collapse in government aid, the funding gaps in these exact areas grew. If you want the highest confidence that a gift did something specific, this is the tier. Our breakdown of the numbers.
Lead exposure. Getting lead out of paint is a regulatory fix that keeps paying off after the charity leaves. LEEP’s own model puts it at a few dollars per healthy year of life, with a 90% range of $0.78 to $12.93. That would make it the cheapest thing on this page. We rank it alongside child health rather than above it because the estimate is a projection, most of the benefit is future income rather than health, and it depends on regulations being enforced. Even at the pessimistic end it is excellent. Full analysis.
Tier 2: probably very good, honestly uncertain
Mental health. Lay-delivered group therapy in Uganda, Zambia and Zimbabwe costs around $45 per person treated and, by HLI’s estimates, produces a year of noticeably better life for $21 to $25. GiveWell’s reading of the same programs is several times less favorable. The truth is probably in between, which still makes it competitive. Where the two evaluators disagree.
Farmed animals. If animal suffering counts, corporate campaigns for cage-free eggs and pre-slaughter stunning of shrimp are among the cheapest ways to reduce it that anyone has costed. ACE’s ten recommended charities and its pooled fund are the practical route. The 2026 picks.
Climate. Policy advocacy groups like the Clean Air Task Force are estimated at under a dollar per ton avoided in expectation, versus $5 to $20 for offsets that often do not deliver. The estimates are old and the politics have moved. Pooled funds run by Founders Pledge or Giving Green rebalance for you. Why the evaluators fund lobbyists, not trees.
Tier 3: the benchmark
Cash. Giving money directly to very poor households works, with the strongest evidence base in development economics and none of the feared side effects. It ranks below the tiers above only because a dollar spent preventing a child’s death buys more welfare, by most people’s values, than a dollar handed to a household. If you distrust every model above, cash is the floor you can trust. The Kenya results.
Tier 4: giving close to home
Most US giving goes to religion, education and local human services. We are not ranking those as bad; we are saying they are almost never measured, and when they are, the cost per outcome is far higher because the cheapest problems were solved in rich countries decades ago. If you give locally for reasons of community or obligation, do it knowingly, and put the share you give for maximum impact into the tiers above.
How we would split a gift
A donor who accepts the evidence and the usual moral weights would put most of an impact-focused gift into Tier 1, a meaningful slice into Tier 2 according to their values (animals, mental health, climate), and use cash as the fallback when a Tier 1 program reports it is fully funded. We do not take a position on how much of your total giving should be impact-focused; that is yours.
What would change this ranking
- GiveWell’s next model update, the first to fully reflect the aid cuts.
- ACE’s November 2026 recommendations.
- Longer-run Kenya UBI data, and results from GiveWell’s cash pilots.
- Any independent replication of LEEP’s model.
We will update this page and note the date at the top when we do.
Method notes
Cost figures are quoted from the evaluators named, with the dates of their analyses. “Evidence quality” and “room for more money” are our judgments, explained in the linked articles. See the methodology for how we compare unlike units.
Sources
- GiveWell: Our top charities
- GiveWell: Cost-effectiveness analyses (funding bar as of May 2026)
- LEEP: How cost-effective are LEEP's paint programs?
- Happier Lives Institute: Our recommended charities
- Founders Pledge Climate Fund: Impact report
- Animal Charity Evaluators: Recommended Charity Fund update, August 2026
- GiveDirectly: Early findings from the world's largest UBI study
Figures come from the cited sources as of the dates shown. Cost-effectiveness estimates change as programs and evidence change; check the source before giving. We are not financial advisers and receive nothing from any charity named here.