We need to talk about carbon credits

Voluntary carbon credits have become big business as more and more companies purchase these offsets to reduce their net emissions while they make structural changes to their operations and try to become net zero in the future. But among ESG investors, these carbon credits are heavily disputed because it is unclear if these schemes really help reduce emissions or avoid deforestation. This can lead to sensationalist newspaper articles like this one.

I have some issues with this and similar press articles. The article claims that 94% of voluntary carbon emission reductions verified by Verra, the world’s largest issuer of voluntary carbon credits should not have been approved because they did not reduce carbon emissions. This claim is based on two factors.

First, Verra may have overstated the estimated reduction in carbon emissions from deforestation by 400% or more. This resulted in Verra selling too many carbon offsets that are in reality not backed by existing forest protection projects. That is a serious claim and if true, Verra has a lot of explaining to do. But I have so far not seen any data that confirms this claim. And while I think the journalists making these claims have this data, I would appreciate it if it could be made available to investors like me to check what is going on there. Again, if Verra or other carbon credit certification organisations sell too many certificates that is a major issue and needs to be addressed urgently.

The second factor that drives the claim that most carbon credits are not really reducing carbon emissions is based on three academic studies on the effectiveness of emissions reduction projects.

All three studies tried to do the same thing but with different methodologies. They looked at the forestation of a series of conservation projects over time and compared it with the forestation (or deforestation) of a suitable control area. If the deforestation in conservation areas is lower than in control areas, the project is considered successful in creating a contribution to reducing climate change. The more trees are protected from deforestation the more carbon credits can be issued against them.

Now, there is an argument to be had if nature conservation should be counted as a positive contribution to reducing greenhouse gas emissions or whether only newly planted trees that would not have been planted otherwise should be counted. But that discussion is beyond the scope of this post.

Instead, let’s focus on these three studies and their results. I read through the two studies of the research group around Thales West published in 2020 and 2022. Both these studies claim that the conservation projects examined (12 Brazilian projects in the 2020 study and 27 projects globally in the 2022 study) did not reduce deforestation compared to a synthetic control group.

Meanwhile, a third study published in 2022 of 40 projects in nine countries came to the conclusion that deforestation was reduced by 47% in the first five years of the projects and degradation of forests was reduced by 58%.

Here is a key chart from the studies of Thales West showing no effect on the rate of deforestation in their studies.

Average reduction in deforestation in one study

Source: West et al. (2022)

And here is the estimated reduction in deforestation and degradation in the other study.

Average reduction in deforestation in another study

Source: Guizar-Coutino et al. (2022)

What is going on here?

The two studies use different methodologies to estimate deforestation in the control group. The group that found a material impact on deforestation used a comparison of satellite images. In essence, they defined a similar area close to the conservation project with a similar degree of forestation and natural conditions. Importantly, the conservation project and the control groups were selected so that they were at a similar distance to recently deforested areas. Then they looked at the pixels in satellite images that showed trees vs. no trees. If trees in satellite images disappeared faster in the control area than in the conservation area, the project was considered a success in reducing deforestation.

The group that found no impact on deforestation instead focused on demarcated land areas as defined by the local land registry. These land areas typically are a mix of protected and unprotected land, and the control groups were designed in such a way that the mix of protected and unprotected land was similar to the mix in the area that contains conservation projects.

I prefer the methodology of the group that found a significant impact. The method of looking at pixels of satellite images is not perfect, but at least straightforward and not too ambiguous. Meanwhile, looking at land areas that are demarcated in the land registry of countries like Brazil and others is rather imprecise. After all, I wouldn’t trust that the Brazilian government has up-to-date and accurate boundaries for some piece of land in the middle of the Amazon rainforest. Furthermore, if you look at land areas that have a mix of protected and unprotected land what could happen is that while the protected land is indeed able to preserve trees at a measurable rate, the unprotected land surrounding it may be deforested more rapidly. After all, people who make money off of deforestation don’t just go away just because some land is suddenly protected. I would expect them to simply go next door and increase their efforts there. That would imply that while the mix of protected and unprotected land starts out as identical in the treated and control groups, the conservation activities in the protected areas are negated by faster deforestation nearby which may result in a net effect for the covered land area of zero.

I do not know if this is what happens here, but it shows that the methodology, with which an effect is measured can influence results widely. And it means that we need to have globally agreed on standards on how to measure the effect of forest preservation.

And going back to the initial point about overstating the carbon reduction projections, we need to have audited and verifiable measures on how much carbon has been saved and how that links to voluntary carbon credits sold to businesses.