Ask most people how to give for the climate and they will say plant trees or buy offsets. The two evaluators that specialize in climate giving, Founders Pledge and Giving Green, almost never recommend either. They recommend small nonprofits that push for policy and technology changes. The reasoning is worth understanding because it generalizes to a lot of giving.
The problem with offsets
An offset is a promise that your dollars caused a ton of carbon dioxide not to be emitted, or to be removed. The promise has to clear three bars: the reduction would not have happened anyway (additionality), it will not be reversed (permanence), and nobody else is counting it too. Forest projects, the most popular kind, struggle on all three. Investigations in 2023 into the largest certifier’s rainforest credits found that most of the credits examined represented little or no real reduction. Prices of $5 to $20 a ton reflect that doubt.
The case for policy
A rule that shuts down a set of coal plants, or a tax credit that makes carbon capture pay, moves millions of tons. If a small advocacy group had even a modest chance of causing that rule, the expected tons per dollar are enormous.
The best-known example is the Clean Air Task Force (CATF), a US group that works on power-sector regulation, methane, advanced nuclear and carbon capture policy. Founders Pledge’s analysis of CATF’s role in the 45Q carbon-capture tax credit put the cost at roughly $0.11 per ton of CO2-equivalent in expectation, and its broader US work at about $0.30 per ton. Its overall range for CATF was $0.10 to $1 per ton. Giving Green independently reaches the same recommendation.
“In expectation” is doing heavy lifting. It means: multiply the tons a policy moves by the probability the group’s work was decisive, then divide by cost. The probability is a judgment call, and a policy can be repealed. A fair reading is “probably a lot cheaper than offsets, with wide error bars,” not “$0.11 is the price.”
Where the money went in 2025
Founders Pledge’s Climate Fund, which pools donations and regrants them, reported $38.8 million granted in 2025 to 15 organizations. CATF has received more than $5.6 million from the fund since 2020. Other 2025 grants it highlights: over $2 million to the EFI Foundation (energy policy research led by former US Energy Secretary Ernest Moniz) and $3.5 million to the Innovation Initiative.
The fund’s picks share a pattern: neglected problems (methane, industrial heat, nuclear licensing, geothermal), leverage over public money or regulation, and a bet that a few million dollars of expertise can steer billions.
The catch
- Timing and politics. US federal climate policy has swung hard in both directions since 2022. Advocacy that paid off under one administration can be undone by the next; the evaluators’ response has been to shift toward technology, state-level and international work, but it lowers the confidence in any single number.
- No feedback loop. With a malaria net you can count deaths. With policy advocacy the counterfactual is unknowable, and the groups being evaluated are also the ones telling the story.
- Crowding. CATF’s budget has grown many times over since the evaluators started recommending it. The marginal dollar is likely less valuable than the average one was.
What to do
If you want a warm feeling and a certificate, buy a high-quality removal credit and accept that you are paying $100 or more per real ton. If you want the most expected tons per dollar, give to a pooled fund run by one of the two evaluators; they rebalance as the politics shift, which you cannot do from the outside. Either way, skip cheap forest offsets.
Method notes
Cost-per-ton estimates are Founders Pledge’s, as summarized by Effektiv Spenden and discussed on the EA Forum; they date from analyses between 2018 and 2021 and have not been updated for recent US policy changes. Grant totals are from the Climate Fund’s 2025 impact report. The offset investigation referenced is the January 2023 joint reporting by the Guardian, Die Zeit and SourceMaterial on Verra-certified rainforest credits.
Sources
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.