Redefining Carbon Credits

Author: Selma Sharaf and Ben Silton

A video has been circulating of climate activist Greta Thunberg walking out of a COP26 meeting on “improving the integrity of voluntary carbon offsets”, referring to the discussions as “greenwashing”. As carbon neutrality pledges become increasingly widespread, carbon offsets are growing in both popularity and contentiousness. 

Offsets (AKA “credits”) are certificates representing the reduction of one metric ton of carbon dioxide from one source (e.g., planting trees) to compensate for the equivalent amount of carbon emitted from a different source (e.g., power plant). Credits offer a convenient solution for companies to meet their net-zero carbon pledges. Indeed, voluntary offset demand is expected to grow up to 100X over the next 30 years.

Source: McKinsey

Organizations such as Verra and Gold Standard offer “verified” offsets, but a lack of robust and consistent standards has led to rightful criticism. If — and that’s a big IF — net-zero becomes increasingly mandatory for certain firms, how will we ensure that physical carbon abatement meets or exceeds what is represented in the offset markets? How can technological tools and ideas be leveraged for this purpose? In this post, we explore challenges in the carbon credit space including additionality, permanence, and double counting, all of which invite crucial innovation over the next few years.


Let’s consider a case in which a company purchases a solar farm to “offset” their emissions. If that farm generates a profit (as solar does), it is likely that some other entity would have bought it regardless of the organization’s actions. In other words, it’s difficult to argue that the company brought new carbon abatement into existence when financing could have easily come from somewhere else. 

In most cases, there is no way to definitively prove whether a project would have been carried out had it not been packaged as an offset. On the other hand, it is often possible to identify projects that are blatantly non-additional. In fact, the Nature Conservancy faced a scandal last year when they sold “meaningless carbon offsets” to companies such as JPMorgan, Disney, and BlackRock, claiming to protect forests that they themselves owned and that therefore could not have been threatened. Similarly, in 2015, the Massachusetts Audubon Society received over 600,000 credits, which they in turn sold for about $6 million in exchange for preserving forests… that were already under their care as a conservation nonprofit. The majority of these credits were purchased by oil and gas companies, enabling them to continue emitting while showing their shareholders and the public “progress” towards carbon neutrality.

These examples of “gaming the system” help demonstrate that true carbon credits should demonstrate at least some “additionality” such that the actions they finance could and would not have been undertaken in the absence of the offset sale.


Many carbon credits use tree planting to compensate for carbon emissions, but there are significant uncertainties that stem from attempting to quantify the carbon uptake over time from trees.

Main carbon stocks and fluxes in forests (Source: Forestry Commission)

If trees are planted as part of a carbon offset project only to be burned down a year later, that forest may be assuming the role of a net emitter rather than a net sink of carbon. Appropriate monitoring and forest management techniques would allow us to quantify carbon capture from forests and make these offsets more meaningful, but tools to enable this do not currently exist at a scale that matches the rising demand for carbon offsets. This challenge is compounded by the fact that different trees offer different amounts of carbon absorption (even when diameter is held constant), as shown in the figure below.

Source: University of Minnesota Extension

An analysis of radar data from a carbon offset project launched in Cambodia in 2008 revealed that only 46% of 13 protected sites were still standing ten years later. Many of the protected areas were logged and converted into villages and military bases. In response to these findings, Virgin Atlantic announced that it would stop purchasing these offsets. However, the project still managed to sell 48,000 credits and remained on the carbon credit market as of 2019.

Range of outcomes at five of the sites from a REDD project in Cambodia (Source: ProPublica)

Double Counting

Carbon accounting is still a nascent industry and is subject to nowhere near the same standards nor scrutiny as financial accounting. The GHG Protocol has been widely adopted by companies as a framework for estimating how much they are emitting, which translates to how much they need to reduce or offset their carbon by in order to achieve their targets. However, a similar set of standards does not exist for calculating the amount of carbon sequestered by an offset project. In general, offset holders must “retire” their offsets in order to claim them and ensure that they are not traded again. When two entities both point to the same credit as compensating for their individual emissions, this is double counting.

Controversy surrounding this issue delayed decisions about the terms of Article 6 of the Paris Agreement for years. If a foreign company purchases a credit to help achieve its carbon neutrality pledge, and that credit is also within the scope of the host country’s contribution to the Paris pledge, there is potential for it to be double counted. Article 6 seeks to address double counting to some extent and was formally adopted at COP26, making it the final article of the agreement to be approved. The next few years will shape the future of the voluntary offset market as the impacts of this long-awaited decision play out.

The Challenge of Addressing Low-Quality Credits

Carbon credits vary in effectiveness and price, and low-quality credits are widespread. These inexpensive offsets present several problems, as outlined by John Sterman, co-director of the MIT Sloan Sustainability Initiative. First, in order for tree planting offsets to exhibit additionality, they must occur on land that would not otherwise be reforested. This land is likely to be more expensive, either due to its value in other uses or because larger quantities of it must be purchased and maintained to compensate for its higher tree mortality rate. In addition, cheap forestry offsets are often monocultures of fast-growing species, but natural forests (though more expensive) offer greater biodiversity and less vulnerability to disease, as well as co-benefits such as supporting tourism jobs and indigenous practices. In fact, researchers have found that forests with multiple tree species may be able to store double the amount of carbon as monoculture plantations. Third, offering inexpensive but ineffective offsets engenders the idea that compensating for carbon emissions is cheaper than it actually is, encouraging companies to purchase more offsets rather than reduce their emissions.

Sustainability departments within large organizations are frequently understaffed and underfunded, if they exist at all. The time investment required to comprehensively assess credits prior to purchase is often unrealistic, and even extensive research is limited by factors that cannot be predicted. For example, a university might decide to replace their diesel buses with electric buses. If the buses are permanently taken out of use, then this credit might be effective; but if another school purchases the old diesel buses secondhand and puts them into service, then there may be no net reduction in carbon. No amount of research can determine how external actors will respond to a well-intentioned project, and “side effects” of organizational decisions are often ignored despite their potential to affect the net carbon abatement of an offset project.

Can Carbon Credits be Sustainable?

Despite these ongoing issues, carbon credits are on the rise. The global carbon offset market is still small, but is expected to reach $50 billion by 2030 and $200 billion by 2050. Over 200 organizations have signed The Climate Pledge, which calls on companies to reach net-zero carbon by 2040. Universities are aiming for carbon neutrality, often as a part of groups such as Second Nature’s Climate Leadership Network. Organizations have decided to adopt offsets as a cost-effective strategy for sustainability, but the concept needs to be reframed. Projects with quantifiable and verifiable effects could help set new standards for carbon credit markets. 

Methane capture and use can be viewed as a potential alternative to forestry offsets. Monitoring the impacts of methane projects is much more straightforward in comparison to tracking all of the trees in a forest and estimating how much carbon they emit and sequester. Methane offset projects exist but currently face a range of barriers, reflected by lower-than-projected quantities of captured emissions.

What if organizations could compensate for their carbon emissions by funding energy-saving retrofits (e.g., heat pump installation) for households in under-resourced communities (with the additional benefit of contributing to a more equitable energy transition)? If the homeowners receiving the retrofits would not have otherwise had access to those improvements (due to lack of affordability or limitations for renters), the additionality of the project can be claimed with more confidence.

How Technology Might Transform the Market

A separate but synergistic approach to addressing the carbon offset issue is advancing technological solutions to improve carbon accounting. This might involve using machine learning and aerial imagery for carbon quantification and post-implementation monitoring of projects. Researchers have identified systematic overestimation of carbon sequestered by forestry offsets. Verification bodies and researchers are working on remote sensing technologies to automate parts of the verification process for these types of offsets. Currently, organizations providing forestry credits tend to manually calculate sequestered carbon using simple conversion factors or tools such as the Winrock calculator. Developing aerial imagery-based ML algorithms would allow for tree species or family classifications to be identified, providing higher resolution data that can be regularly updated for monitoring purposes.

Another pathway is using big data tools to improve carbon offset databases. This could involve consolidating offsets from various registries into a single directory, improving transparency and making information more accessible. Initiatives such as Berkeley’s Voluntary Registry Offsets Database and AlliedCrowds’ AlliedOffsets are being developed within this space, and offset databases are beginning to integrate natural language processing (NLP) and other scalable classification methods. This could enable the databases to extract more information about each offset, including whether anyone has been tasked with monitoring the project and how the verification process is being carried out.

Clean blockchain (from renewable energy sources) can also be used to create transparency in the market by replacing traditional accounting methods. Firms looking to blockchain for carbon offsetting include Climatetrade and Everledger. This solution is attractive because information entered through blockchain is immediately distributed across a network, providing traceability and helping ensure that credits are not double-counted or resold.

These are only a few of the technological opportunities for transforming the carbon credit market. There is an urgent need to redefine the concept of carbon credits, and innovative ideas must be deployed to meet the rising demand for offsets and address the current challenges. We at ADL would be happy to discuss how your organization might integrate reliable carbon offsets into your larger low-carbon strategy.

Special thanks to Timothy Gutowski (MIT MechE) and Harvey Michaels (MIT Sloan) for their insight on this topic.