Buyer beware: a taxonomy of the risks of international carbon credits

As the EU reintegrates international carbon credits into its climate framework, understanding the associated risks is essential.

For more than a decade, international carbon credits attracted little attention in the European Union. Two related developments changed this. First, the United Nations climate change conference in Baku, Azerbaijan, in November 2024 finalised the rulebook for implementing Article 6 of the Paris Agreement, establishing the framework for the international exchange of carbon credits between participating countries. Second, EU policymakers agreed a climate target for 2040 of a 90 percent emissions reduction, compared to a 1990 baseline. This allows for up to 5 percent of the emissions reduction target to be achieved using international carbon credits issued under the Paris Agreement mechanisms or from domestic projects.

The economic rationale for this new allowance is straightforward: because mitigation costs differ across countries, allowing emissions reductions to occur where they are cheapest lowers the overall cost and makes it more likely climate objectives will be achieved.

Under the amended EU Climate Law (Regulation (EU) 2021/1119), updated to include the 2040 target, 5 percent of emissions equals at least 236 million tonnes of carbon dioxide equivalent (CO2e) between 2036 and 2040 – roughly equivalent to France’s annual CO₂ emissions and worth between €7 billion and €24 billion at a carbon price range of €30/t to €100/t. This does not count potential buyers beyond the EU.

In this context, this analysis maps the fragmented landscape of international carbon crediting, shows that many of the methodological and governance challenges that undermined confidence in Kyoto-era carbon credits remain unresolved under Paris Agreement mechanisms, and develops a classification of the key integrity risks for global climate mitigation. If not monitored carefully, the EU’s introduction of international credits risks increasing global carbon dioxide emissions (Calel et al, 2025).

The Paris Agreement has not fixed the core problems

The 1997 Kyoto Protocol established the first specific commitments for individual countries to limit greenhouse gas emissions and introduced international ‘flexibility mechanisms’. The Clean Development Mechanism (CDM) allowed developed countries to earn credits from mitigation projects in developing countries, while Joint Implementation (JI) enabled earning of credits in other developed countries.

In practice, however, international crediting proved vulnerable to significant integrity risks.  Between 2008 and 2012, companies used international credits to offset around 1.5 billion tonnes of domestic CO2e under the ETS (Delbeke and Vis, 2019). The influx of international carbon credits during this period contributed to weakening the carbon price signal and reduced incentives for domestic decarbonisation (Marcu et al, 2025). In response, the EU progressively restricted the use of Kyoto credits and excluded them entirely from Phase IV (2020-) of the emissions trading system.

However, evidence suggests that neither the Kyoto mechanisms nor the fragmented voluntary carbon market delivered real and additional emissions reductions. Despite efforts to strengthen governance and methodologies, concerns about additionality and emissions accounting remain widespread (Balmford et al, 2023; Macintosh et al, 2025; Romm et al, 2025; Schneider et al, 2024; Schneider and La Hoz Theuer, 2019). 

The Paris Agreement created two market-based cooperation mechanisms. First, under Article 6.2, countries can trade carbon credits directly with each other. This allows one country to count emissions reductions achieved in another country towards its own climate targets. Second, Article 6.4 establishes a UN-supervised global carbon market, known as the Paris Agreement Carbon Market (PACM). Instead of countries trading directly, a central body oversees the system by approving project methodologies, issuing carbon credits and monitoring transactions to ensure environmental integrity.

Despite recent progress in operationalising Article 6, three challenges continue to cast uncertainty over its effectiveness.

Under Article 6.2, responsibility for ensuring credit quality rests directly with buyers and sellers, with no independent supervision. Thus, bilateral agreements may not provide sufficient safeguards for transparency, reporting and environmental integrity. Evidence from Switzerland illustrates this. To meet its 2030 emissions target, it plans to rely partly on international carbon credits, backed by a CHF 2.5 billion budget for 2021-2030. It has entered into bilateral agreements with around 20 host countries to finance mitigation projects in exchange for international carbon credits. The first project, to support the electrification of public buses in Bangkok, has raised integrity concerns, as the bus fleet may have been electrified even without Swiss finance. Limited disclosure of project-level information has also raised transparency concerns.

Second, key issues plaguing legacy credits, such as additionality, permanence and measurement, remain too broadly defined to guarantee robust emissions reductions under both Articles 6.2 and 6.4 (Johnstone et al, 2025). 

Third, concerns remain regarding the effectiveness of the new framework’s environmental integrity safeguards, especially with the transition of around 1500 Clean Development Mechanism (CDM) projects into the new Article 6.4 PACM. In some cases, potentially problematic legacy methodologies will be retained by projects for calculating baselines and emission reductions (Johnstone et al, 2025). Most transferable CDM projects are clustered in Southern and Eastern Asia (Figure 1), with frequent criticism for non-additionality (Cames et al, 2016; Wyburd, 2024).

Figure 1: Potential emissions reductions from Kyoto credits eligible for transfer to PACM and share of total eligible Kyoto credit reductions by region and project type

^{Source: Bruegel based on UNEP. Note: Kyoto credits eligible for transfer refer to Certified Emissions Reductions (CERs). ‘Other’ comprises agricultural, forestry and fuel-switching projects. The percentage indicators show the share of projects coming from single selected countries within regions.}

A taxonomy of environmental integrity concerns

Environmental integrity is a high priority for policymakers and potential buyers of carbon credits. We classify the main risks that can undermine the credibility of carbon crediting projects. A thorough understanding of the risks associated with the use of international carbon credits is essential when attempting to identify ‘high-integrity’ credits, as stipulated by the EU Climate Law. While these risks arise in both domestic and international markets, they are often more pronounced in cross-border crediting arrangements, for which monitoring, verification and enforcement are more complex and information dispersed more asymmetrically. International carbon crediting projects have been shown to often fall short against one or several of the integrity criteria shown in Figure 2 (Macintosh et al, 2025). We deal with each of these issues in turn.

Figure 2: Taxonomy of problems

^{Source: Bruegel.}

Measurement

Estimating emissions reductions from carbon credit projects requires imperfectly measured emissions outcomes to be compared to unobservable counterfactuals of what emissions would have been without the project. This is a persistent challenge, particularly because abatement potential varies widely across project types (Macintosh et al, 2025). 

These measurement difficulties are particularly pronounced in nature-based and waste-gas projects and initiatives to replace open fires for cooking with cleaner alternatives (so-called ‘cookstove’ projects). In nature-based projects, uncertainty over carbon stocks and sequestration rates complicates baseline setting (Macintosh et al, 2025; West et al, 2023). Quantification of baseline emission reductions for renewable energy projects is more straightforward as data on electricity generation and grid-emission factors is often widely available. Cookstove projects often rely on unverifiable assumptions about household behaviour and fuel use, leading to substantial over-crediting in some cases (Gill-Wiehl et al, 2024). Waste-gas projects have similarly been criticised for inflated baselines and excessive credit issuance (Cames et al, 2016; Schneider et al, 2024; Schneider and Kollmuss, 2015) (Table 1).

Permanence

Emission reductions can be non-permanent: credited reductions or removals are reversed over time. This risk is particularly pronounced in nature-based projects, including forest conversation, afforestation, reforestation and soil carbon sequestration, because stored carbon can be released through wildfires, droughts, pests or changes in land use and forest management (Macintosh et al, 2025; Probst et al, 2024).

The claimed emissions reductions from cookstove projects, for example, depend on the long-term preservation of forest carbon stocks, making their permanence contingent on the same risks that affect nature-based projects (Table 1).

Additionality

Additionality is one of the most frequently raised integrity concerns in carbon crediting. At the project-level, a project is ‘additional’ only if it would not have been implemented without carbon credit revenues (Schneider and La Hoz Theuer, 2019). Assessing additionality therefore requires observed outcomes to be compared against a counterfactual business-as-usual scenario. Renewable energy projects have frequently been criticised on these grounds, as many would likely have been developed regardless of carbon-credit financing, particularly as technology costs have fallen and deployment has become commercially viable (Probst et al, 2024; West et al, 2023) (Table 1).

Table 1: Integrity risk by project type

	Nature-based	Renewable energy	Energy efficiency	Waste gases
Measurement	High	Low	High	Medium / High
Permanence	High	Low	Low / Medium	Low
Project  additionality	Low	High	Low	Low / Medium

^{Source: Bruegel.}

At the macro-level, additionality can be undermined by leakage, whereby emissions reductions in one location lead to emissions increases elsewhere. Forest conservation projects illustrate this risk: protecting one area may simply shift deforestation to another region if underlying demand remains unchanged (Romm et al, 2025).

Double counting

Double counting occurs when both the buyer and seller country claim the same emissions reduction towards their climate goals, undermining the global mitigation benefit. Article 6 addresses this risk through “corresponding adjustments”: when a host country transfers credits internationally, it must add the transferred mitigation outcome back to its own emissions balance, allowing the buyer country to count it instead (Romm et al, 2025).

Fraud

Fraud is a transversal issue that can emerge at various stages affecting quality, quantity, additionality, permanence and underlying methodology to assess the abatement potential of credited projects. Profiteers of carbon crediting projects may, for instance, be incentivised to artificially inflate the emissions-reduction potential of credited projects in order to maximise the volume of carbon credits sold or to provide false information relating to the additionality of a credited project, which is generally difficult to prove. Significant information asymmetries between project developers, verifiers and buyers can make it difficult to assess the credibility of project claims. This can create perverse incentives for market participants to misrepresent project outcomes, increasing the risk of fraud (Macintosh et al, 2025). Limited information and low perceived market integrity raise the cost of assessing credit quality, increasing the barrier to market participation.

The EU should not repeat past mistakes on international carbon credits

International carbon crediting remains a highly fragmented activity, characterised by diverging standards and methodologies, while the Paris Agreement’s Article 6 framework still lacks safeguards to ensure that credited emissions reductions are real, additional and permanent.

The EU should therefore treat the evolving Article 6 framework as a starting point – not a guarantee – when designing the role of international carbon credits in its post-2030 climate architecture. Integrity risks at both the systemic and project level can undermine the climate benefits of international credits. 

At the same time, stringent integrity requirements may constrain the rapid scaling-up of credit supply precisely as the EU approaches the 2030 pilot phase. Policymakers therefore face a fundamental trade-off: ensuring environmental integrity in a highly dynamic and fragmented carbon market while securing a sufficient supply of high-quality carbon credits within a constrained timeframe and budget.

Source : Bruegel