When Giving Green started, we were focused on looking for evidence-based climate solutions. But eventually, we came to the conclusion that to maximize our impact, we have to support strategies that are less measurable but can bring about grand, systematic changes.

It started from a place of frustration. One day, feeling especially stressed about climate change, I decided I wanted to do something and make a donation to help the cause. But as an economist working on evidence-based policy, I wanted to make sure that I donated to a cause that had strong evidence to support it. So like anyone, I turned to Google, trying to figure out what climate interventions are really backed by evidence.

Much to my chagrin, this turned out to be really hard. A lot of interventions claiming to offer clear emissions reductions seemed shakier once you scratched below the surface. As far as I could tell, at the time I couldn’t find any centralized source (for instance, Givewell in international development) that was vetting high-impact climate-giving opportunities. So I started Giving Green in 2019 to fill this niche. My quest: to find the most evidence-backed climate solutions and promote them to the world. 

There seemed to be an obvious place to start: the carbon offsets market. This is because carbon offsets are supposed to offer a way for buyers to contribute to projects that are causing verified emissions reductions. But as exposed time and time again, many carbon offset projects are not delivering the emissions reductions that they promise. So in our search for evidence-based approaches, we started by analysing the carbon offset market, trying to identify offsets that truly were delivering the emissions reductions they were promising. 

It wasn’t so easy. It turns out that this market that is supposed to deliver ‘certainty’ to buyers is rife with uncertainty and overpromises. But we found some products that we liked. For instance, although there are mixed results on the effectiveness of cookstoves on carbon emissions (generally due to lack of usage), there was a great randomized controlled trial showing terrific performance of a cookstove manufactured by BURN, who used carbon offsets to subsidize some of their stove sales. In fact, in the original version of our website, BURN was our top recommendation. 

But as we continued assessing carbon offsets, doubts started to creep in. The energy transition we need to combat climate change is vast and complex and requires massive shifts in behaviour and technology. Could you really make progress with a blanket of cookstove projects? Besides our internal doubts, we were getting criticism from colleagues in the climate space. Many forward-looking climate philanthropists believed that the offset market was nothing more than a set of band-aid solutions and were instead focused on making grants that focused on systems change, like policy advocacy. We were being dismissed as a climate evaluator who was missing the forest for the trees.

But for us, making a change was daunting. Our entire reason for existence was to fill a gap in evidence for climate giving, and we wanted our recommendations to be supported by hard, incontrovertible proof. And this kind of evidence was only possible for local projects in which one could carefully measure the project’s impact with data. If we shifted away from this core identity, what would we be left with?

To grapple with this issue, we turned to a favourite tool from our toolkit: cost-effectiveness modelling. It was relatively straightforward to model the cost-effectiveness of a carbon offset in terms of dollar per ton of CO2 (for instance, see our analysis on the cost effectiveness of BURN), this was a lot tougher when considering organizations doing policy advocacy or technology promotion. But we tried our best. For instance, we built a model that estimated reductions in greenhouse gasses from various policies being considered in the US in 2021-2022, many of which eventually were included in the Inflation Reduction Act of 2022. We then made estimates of the influence that various organizations had on specific laws being included in the bill, or of a climate bill passing in the first place. (For instance, see our activism cost-effectiveness analysis for estimates of the effectiveness of ‘outsider’ climate activism in promoting big climate bills to be passed in 2021-2022.) What we found was quite illuminating. Even when making quite conservative assumptions on difficult parameters in the models (such as the amount of influence a certain organization had), advocacy organizations were looking much more cost effective than even best offsets. 

At this point, there was no choice left but to switch our approach. In 2020 we added advocacy organizations to our list of top recommendations, and over time came to strongly believe that the most bang for your buck in climate philanthropy comes from funding organizations working to make systemic change policy and technology, and all of our top recommendations in 2023 work in policy and/or technology advocacy. When making the switch, we did indeed have to lose some of our identity as having really evidence-backed recommendations and had to instead start describing them as the best bets. But we felt it was the right thing to do. If the best bets were ones where we were going to have to embrace uncertainty, that’s just how it had to be.

Growing an organization is hard, and there are always shifts and bumps along the way. We’ve chosen to highlight this shift in strategy along with other errors we’ve made on our journey on our public mistakes page. We’re sure this page will continue to grow, but hopefully not too fast.

Dr Dan Stein is Giving Green’s founder and director, and the chief economist at IDinsight.
Email: givinggreen@idinsight.org