1.1 Relevance of Carbon Markets

How to address climate change is one of the greatest global governance problems of our time. At the international, national, and subnational levels, over 50 different carbon markets have been implemented as a key policy to incentivize the reduction of greenhouse gas (GHG) emissions in a cost-effective and flexible way, and several more are being planned or considered (World Bank, 2021). In addition, 89 Nationally Determined Contributions (NDCs) submitted by Parties to the 2015 Paris Agreement mention the use of carbon markets as a condition for achieving their mitigation targets (Reference Pauw, Cassanmagnano and MbevaPauw et al., 2016). Article 6 of the Paris Agreement envisages the implementation of carbon markets or similar international cooperative arrangements as a policy instrument to facilitate the achievement of its goals.

Carbon markets are markets where a certain amount of GHG (e.g., a tonne CO₂ equivalent) is commodified as a tradable unit either as an emission allowance issued under a cap-and-trade system or as a verified emission reduction/removal credit issued under a baseline-and-credit system.

In “cap-and-trade” or emissions trading systems (ETS), a regulator defines an allowed maximum level of GHG emissions (the “cap”) for a certain group of entities (e.g., countries, companies, or facilities). The cap is then subdivided into distinct emission allowances, which are distributed to the regulated entities. The covered entities need to submit one allowance for each tonne of carbon dioxide equivalent (CO₂e) emitted during a compliance period, usually a year. The initial allocation of allowances to covered entities can be free of charge, e.g., based on historical emissions levels (“grandparenting”), partially free (with free allocation limited by a politically determined technology performance benchmark), and/or sold at auction by the regulator.

In a “baseline-and-credit” system a regulatorFootnote ¹ defines how emission (reduction or removal) credits can be generated by activities that reduce GHG emissions or remove GHGs from the atmosphere compared to a reference scenario (baseline) that reflects the counterfactual situation without these activities. The difference between the baseline emissions and the emissions of the activity determines how many credits can be issued. To generate emission credits, ex post verification of the reduction/removal by an officially recognized institution – a verifier – is necessary. The emission credits can then be used as offsets against mandatory or voluntary GHG emission targets or other policy instruments aiming at GHG mitigation. Table 1 shows the key differences between a baseline-and-credit and a cap-and-trade system.

Both types of units form the supply in the market. There can be different types of demand for allowances or credits at different levels. Governments can use units to comply with emissions targets under an international treaty such as the Kyoto Protocol or the Paris Agreement. Companies can use allowances to comply with their targets under emissions trading schemes. In some jurisdictions, they can use credits in emissions trading systems, dedicated baseline-and-credit systems for specific sectors, or instead of having to pay carbon taxes (e.g., as allowed in Colombia or South Africa). Finally, private companies and individuals can use credits for offsetting emissions in the context of their voluntary GHG mitigation targets; such demand has increased significantly in the past years.

The carbon price is discovered in both compliance and voluntary markets through the buying and selling of units, whereas the scarcity of units and the marginal costs of reducing greenhouse gases influence the price. The initial allocation or issuance of allowances and credits by the regulatory authority represents the primary carbon market. Allowances and credits can then be traded in the secondary carbon markets (spot market), either directly between parties, usually facilitated by brokers (over-the-counter transactions, OTC), or traded on an exchange. While the latter requires prior standardization of contracts, for OTC transactions, the transacting parties can freely shape the contract in terms of, for example, price and volume of units being traded. Since the details of these contracts are generally not published, OTC transactions can be quite opaque for other market players and regulators (Reference Kachi and FrerkKachi and Frerk, 2013). A further component of carbon markets is the derivative market. It is composed of financial instruments, such as options and futures contracts, to hedge the risks associated with emission allowances and credits.

We would like to note that, in practice, the terminology is not always used consistently. While the IPCC Assessment Reports (Reference Gupta, Tirpak, Burger, Metz, Davidson, Bosch, Dave and MeyerGupta et al., 2007; Reference Stavins, Zou, Brewer, Edenhofer, Pichs-Madruga and SokonaStavins et al., 2014) and all relevant carbon market research literature (e.g., Reference Michaelowa, Shishlov, Bofill, Hoch and EspelageMichaelowa et al., 2019b) apply the terms as defined in this volume, a few practitionersFootnote ² further differentiate baseline-and-credit systems into those that use emission credits for offsetting and those (national or subnational) systems in which baseline emission levels are defined for individual regulated entities (e.g., based on historical levels or on an industry standard) and units are issued to entities that have reduced their emissions below this level. Under such a system, units can be sold only to other entities exceeding their baseline emission levels. Following the research literature, in our volume we use the term “baseline-and-credit-system” in a broad sense, covering all types of markets in which emission credits are issued compared to a baseline. We also note that occasionally the literature uses the term “emissions trading” as an umbrella for all systems described above.

1.2 Carbon Markets Around the World

Currently, at least 29 ETS and 27 baseline-and-credit systemsFootnote ³ are in place around the world, covering international, supranational, national, and subnational jurisdictions (World Bank, 2021).

At the global level, the 1997 Kyoto Protocol to the UN Framework Convention on Climate Change (UNFCCC) introduced three market-based flexibility mechanisms: the Clean Development Mechanism (CDM), Joint Implementation (JI), and International Emissions Trading (IET). The CDM is a baseline-and-credit system that finances emission reduction projects in countries without emission reduction targets under the Kyoto Protocol (so-called non-Annex I countries). The Certified Emission Reductions (CERs) generated by these projects can be used by countries with targets under the Kyoto Protocol (Annex B countries) toward their own compliance. JI is a similar baseline-and-credit mechanism, which operates in Annex B countries. The units it generates are called Emission Reduction Units (ERUs). There are two forms of JI: Track 2, which is subject to international oversight, and Track 1, which is not. IET allows Annex B countries to trade the Kyoto Protocol’s unused Assigned Allowance Units (AAUs) with each other.

Under its Article 6, the 2015 Paris Agreement specifies the implementation of similar market-based mechanisms, with detailed rules agreed by COP26 in 2021. Direct bilateral cooperation under Article 6.2 allows, for example, the linking of national, subnational, and supranational ETS and the trading of so-called Internationally Transferred Mitigation Outcomes (ITMOs) in a way comparable to IET and to JI Track 1 projects. A multilaterally overseen Article 6.4 Mechanism will be a baseline-and-credit system similar to CDM and JI Track 2 (Reference Michaelowa, Shishlov and BresciaMichaelowa et al., 2019c). In addition, in 2016, the International Civil Aviation Organization (ICAO) established a pilot baseline-and-credit mechanism known as CORSIA (Carbon Offsetting and Reduction Scheme for International Aviation). CORSIA, which started operating in 2021, aims to incentivize carbon-neutral growth of the international aviation sector.

At the supranational level, the EU ETS is the largest ETS currently in place. It covers installations in the power and heat generation, energy-intensive industry, and commercial aviation sectors in all 27 EU member states as well as Iceland, Liechtenstein, and Norway. It has set emissions limits for more than 11,000 installations and airlines, covering about 40 percent of the EU’s GHG emissions.Footnote ⁴

So far, eight national-level ETS are operating, and more are being planned. Subnationally, several Canadian, Chinese, Japanese, and US jurisdictions have implemented or are planning ETS. In addition, several of these jurisdictions have implemented baseline-and-credit systems to supply offsets to their ETS (ICAP, 2021; World Bank, 2021).

Figure 1 presents a simplified overview of the main international carbon markets, their linkages, and traded units.

Figure 1 Examples of trading systems, linkages and traded units

Source: Own graphic. The red arrows depict direct links between the different systems which allow them to trade with each other. In addition to this, some markets may be linked indirectly – e.g., the EU ETS and the NZ ETS are connected through their link to the Kyoto Protocol mechanisms.

This volume does not aim to be comprehensive and cover all carbon markets but rather focuses on major compliance markets. For a comparison of national-level baseline-and-credit schemes see Reference Michaelowa, Shishlov, Bofill, Hoch and EspelageMichaelowa and colleagues (2019b), while an excellent overview of national emissions trading schemes has been done by Reference HaitesHaites (2018). It also does not provide a comprehensive description of the political processes that led to the evolution (and improvement) of carbon markets over time (for this, see, e.g., Reference Wettestad and JevnakerWettestad and Jevnaker, 2016, Reference Wettestad and Jevnaker2019; Reference Michaelowa, Shishlov and BresciaMichaelowa et al., 2019c). Rather, its goal is to focus on the risks that surround the design of carbon markets and the solutions that have been devised to address those risks. Instead of offering full case studies of individual carbon markets, then, we use them to derive lessons across the various design aspects of carbon markets.

1.3 Carbon Markets as Polycentric Governance Arrangements

Carbon markets are complex governance arrangements (see Reference Ahonen, Kessler, Michaelowa, Espelage and HochAhonen et al., 2022). They entail artificially created markets for goods – emission allowances or credits – that are created by policy. As described above, they have emerged at all levels of governance, and engage public and private actors in regulatory and governance functions. They comprise various, mostly independent governance systems that are interlinked – in a few cases through formal market links, but in most cases loosely through flows of information and expertise, capacity building, and through the overarching goal of helping to achieve international climate mitigation commitments (Reference Burtraw, Palmer, Munnings, Weber and WoermanBurtraw et al., 2013; Reference Paterson, Hoffmann, Betsill and BernsteinPaterson et al., 2014; Reference Biedenkopf, Müller, Slominski and WettestadBiedenkopf et al., 2017). For these reasons, carbon markets are inherently polycentric in nature, involving multiple and often overlapping sources of authority (Reference Jordan, Huitema, van Asselt, Forster, Jordan, Huitema, van Asselt and ForsterJordan et al., 2018).

When the sources of authority cross the boundaries between different levels of public and private governance, the literature speaks of transnational governance. As early as in 2008, the CDM was portrayed by Reference Pattberg and StripplePattberg and Stripple (2008) as a prototype of this form of governance; the Article 6.4 mechanism under the Paris Agreement is comparable. At the UNFCCC level, the parties adopt political guidance, whereas the Supervisory Body takes care of day-to-day decision-making. At the national level, state-appointed Designated National Authorities approve potential Article 6.4 activities in the host country. Private entities – so-called Designated Operational Entities (DOEs) – validate and verify the activities and their baseline methodologies (see Figure 2). Further private actors are involved in identifying, proposing, and investing in activities. Some of them also help to shape the rules and regulations that govern this market. Finally, multilateral organizations, including through public–private partnerships, provide technical advisory, capacity building, and project finance. Under the CDM this was done by the World Bank’s Prototype Carbon Fund; more recently, the World Bank Partnership for Market Readiness took up a similar role.

Figure 2 Activity cycle under the Article 6.4 mechanism

Source: Own graphic.Note: A6.4ERs = Article 6.4 Emission Reductions (tradeable units under Art. 6.4); SOP = Share of proceeds (tax for financing adaptation and administration). Colors denote the actors involved in the process.

The EU ETS involves much authority at the supranational level, along with the national level, with a deep intertwining of decision-makers at both levels, and an additional involvement of private actors (Reference Bailey and MareshBailey and Maresh, 2009). The balance of authority exercised by the supranational and the national levels – for instance, in terms of allocation of allowances – has important implications for the environmental effectiveness of the system (Reference ClòClò, 2009).

In addition, the voluntary carbon market represents an extreme case of involvement of private actors as sources of governance. In words still fully adequate today, Reference Pattberg and StripplePattberg and Stripple (2008: 378) described the voluntary carbon market as “a site of climate governance beyond the state,” in which various, mostly private, actors compete in developing validation and verification standards and providing offsets of varying qualities. Over the last years some large voluntary carbon market standards such as Gold Standard and partially Verra have become more stringent than international compliance markets like the CDM, while small niche standards have offered low-quality credits (Reference Acworth and KunemanAhonen et al., 2022).

Governance arrangements for carbon markets, therefore, exemplify a much broader trend in global environmental governance, away from traditional state-centered multilateral regimes, and toward a co-existence with private and hybrid forms of governance in which new types of actors exhibit authority, experiment with novel arrangements and policy instruments, and interact at various jurisdictional levels (Reference Pattberg and StripplePattberg and Stripple, 2008; Reference Biermann, Pattberg, van Asselt and ZelliBiermann, 2010). In the past decade, these complex multilevel, and polycentric, arrangements have been extensively researched in terms of their implications for effectiveness, institutional interaction, policy learning, diffusion, and convergence, as well as the role of local cultures, ideology, and political economic factors in shaping them (for the case of carbon markets see, e.g., Reference ClòClò, 2009; Reference WettestadWettestad, 2009; Reference Knox-HayesKnox-Hayes, 2016; Reference Biedenkopf, Müller, Slominski and WettestadBiedenkopf et al., 2017; Reference Wettestad and GulbrandsenWettestad and Gulbrandsen, 2018).

These novel arrangements entail opportunities but also risks for climate governance and its effectiveness. In terms of opportunities, carbon markets are characterized by a strong degree of policy experimentation and learning (Reference Biedenkopf, Müller, Slominski and WettestadBiedenkopf et al., 2017), which can lead to solutions that will, over time, be better for the climate. At the same time, given the increasingly decentralized (or “fragmented,” see Reference Biermann, Pattberg, van Asselt and ZelliBiermann et al., 2009; Reference Biedenkopf, Müller, Slominski and WettestadBiedenkopf et al., 2017) authority, it is likely that these arrangements will lead to insufficient, patchy, and uncoordinated regulation that may result in regulatory loopholes and abuses. Misaligned interests and resources between the different sides involved in policymaking may further increase these risks (see, e.g., Reference Bailey and MareshBailey and Maresh, 2009).

It is, therefore, crucial to take a closer look at how these governance arrangements perform in terms of regulatory quality, and what risks they entail for ensuring environmental integrity and improving economic efficiency.

Contrary to what policy diffusion theory would predict, there has been a high degree of divergence in market design as policymakers have learned from the mistakes of previous experiences, and as they have adapted existing designs to the structural and political particularities of their own jurisdictions (Reference Biedenkopf, Müller, Slominski and WettestadBiedenkopf et al., 2017; Reference Wettestad and GulbrandsenWettestad and Gulbrandsen, 2018). This has led to great variation in the regulatory and design aspects of carbon markets (Reference Gulbrandsen, Wettestad, Victor and UnderdalGulbrandsen et al., 2019).

In a context in which we now have substantial experience with various carbon markets at different levels, but where many new markets are being planned, there is a need to survey these various market design characteristics, identify where there are risks for environmental integrity and economic efficiency, and propose lessons for future policy design, with a particular emphasis on sound regulatory oversight. That is what this volume seeks to achieve.

2.1 Legal Principles for Carbon Markets

Generally, principles are variously understood as a “fundamental truth or doctrine,” “a proposition so clear that it cannot be proved or contradicted unless by a proposition which is still clearer,” “that which constitutes the essence of a body or its constituent parts,” and “that which pertains to the theoretical part of a science” (Reference BlackBlack, 1990: 1193). Their scope, therefore, is vague and abstract. In a legal context, however, principles acquire a narrower, more formal role and can take on specific legal effects.

Principles are an integral part of most legal systems and considered necessary for their functioning, without necessarily being set out in written law (Reference Kohen, Schramm and CartyKohen and Schramm, 2013). Their normativity may derive from long-standing practice and legal custom or from accepted requirements of fairness and justice (Reference DworkinDworkin, 1978). Sometimes, however, legal principles are also expressly set out in treaty instruments, statutory law, or even constitutions.

Although the existence of legally relevant principles is hardly disputed, their exact role and definition remain a matter of continued jurisprudential debate. Their legal scope is generally acknowledged to be general and abstract. They apply to a broad and unspecified set of actors and situations which can include future and as yet unknown circumstances or transactions.

They are thus distinct from rules, which are more binary in nature and call for an automatic outcome whenever specified conditions are met (Reference DworkinDworkin, 1978). Principles do not involve such an automatism. Rather, their role is often seen as subsidiary. They guide administrative discretion or judicial decisions where legal rules allow for different outcomes, due to textual ambiguity, the existence of substantive gaps, or conflicts between different applicable rules (Reference Kohen, Schramm and CartyKohen and Schramm, 2013). As a result, legal principles rarely have the normative force to decide a dispute or determine a legal question on their own. They tend to require further elaboration through legislation, case law, or scholarly writing to take effect.

Several legal principles may be relevant for the regulation of carbon trading. It would be difficult – if not impossible – to enumerate all of them, however. Across legal systems and jurisdictions, and even across different areas of law, countless principles of varying weight and degrees of conceptual clarity exist, with sometimes subtle differences in terminology and material substance. We consider selected key principles of public law, private law, and the cross-cutting substantive area of environmental law to potentially be the most relevant ones.

2.2 Environmental and Economic Principles for Carbon Markets

In an environmental and economic context, principles are more widely understood as a “moral rule or standard of good behavior,”Footnote ⁵ which is distinct from the more formal legal principles described in the previous section. In this section, we therefore define and describe the main environmental and economic principles that will guide our assessment of carbon markets and lay out some design elements that have an impact on those principles. Table 3 lists and describes these elements and their potential impact and points out to the sections of this volume where they are addressed.

3.1 Characteristics of Baseline-and-Credit Systems and Possibilities for Abuse

Two main criteria are typically considered when assessing the quality of a baseline-and-credit system: additionality and setting an appropriate baseline. Both are key to ensuring environmental integrity, and both, therefore, offer the greatest risk for abuse and damage to the environment.

“Additionality” refers to whether a project, for instance an investment in a more efficient production process, is actually mobilized by the revenue coming from emission credit sales, or whether the investment would have happened anyway. If a project is viable even without the incentive of selling credits, then it is not additional. If non-additional activities are accepted under a mechanism, the cost of credit generation will be zero or even negative. Players with non-additional activities can, therefore, sell credits at any price that covers the transaction costs for credit generation. In this case, the credit volume will increase while the price will decrease.

The other main risk for abuse is related to how baseline emissions levels are set (Reference Michaelowa and YaminMichaelowa, 2005). Abuse in the context of baseline setting means that baseline emissions levels are set at a higher level than would be the case if the baseline had been specified in line with environmental integrity. Such abuse can be undertaken by governments or activity developers and happen in different ways, including through the use of manipulated data or by increasing emissions beforehand to achieve higher reductions later (see Section 3.2.1). Compared to a situation with correct baselines, the volume of credits will increase. If the credit demand does not change, the equilibrium price for credits will decrease.

In both cases, the extra credit volume generated does not represent a real reduction in emissions. If the credits are used for compliance under a cap-and-trade system, the stringency of the cap will de facto be reduced by this extra credit volume and more emissions will be released into the atmosphere compared to a situation without the possibility of surrendering credits.

Regulators have long been aware of the risks of baseline and additionality abuse and have applied different approaches to address them. Under the Kyoto mechanisms CDM and JI, as well as the voluntary carbon market standards Gold Standard and Verra, validation of project documentation by third-party auditors is mandatory. This validation aims to provide an independent audit and to uncover cases where data used to argue additionality are manipulated. Auditors need to be accredited with the regulators to ensure that their quality is sufficient. However, the fact that validators have an interest in being hired again by project/program developers generates a disincentive to check documentation too carefully. If a validator rejects a validation, the program developer is less likely to hire the validator in the future. This disincentive is particularly strong if the project development is undertaken by highly specialized companies, as was the case under the CDM. A similar disincentive applies for verification of emission reductions achieved by projects, especially if verifications are frequent. These disincentives could be eliminated by random allocation of validators and verifiers to projects by the regulators. Such an approach would require the definition of a common fee scale for validation and verification to prevent excessive fees. The New South Wales Greenhouse Gas Abatement Scheme (GGAS), which ran from 2003 to 2012, followed this approach, requiring that the regulating authority was the client of the third-party verifiers and validators.

There is a surprisingly small amount of academic literature empirically addressing these risks. Reference Drew and DrewDrew and Drew (2010) provided a first rough general overview of fraud under the CDM. Only very recently, Reference Chen, Letmathe and SoderstromChen and colleagues (2021) have published a highly elaborate empirical study assessing over 2000 CDM projects regarding additionality fraud. This study is discussed in detail below.

3.2 Risks Related to Baseline Setting

Baseline methodologies differ significantly across baseline-and-credit mechanisms. The Kyoto mechanisms CDM and JI have approved over 250 methodologies, while other mechanisms, such as the Joint Crediting Mechanism (JCM), and mechanisms under voluntary market standards, such as the Gold Standard and Verified Carbon Standard (VCS), add another 50 methodologies.

Under the CDM, three principal approaches for baseline setting were applied: historical emissions levels, emissions levels of an economically attractive course of action, and the best 10 percent of comparable technology (benchmark approach).

In many baseline methodologies, the correct estimation of the duration during which a technology is effectively used within a typical week (or month) plays an important role. Overestimation of usage duration increases baseline emissions and generates upwardly biased estimates of emission reductions.

Reference Rosendahl and StrandRosendahl and Strand (2009) and Reference Strand and RosendahlStrand and Rosendahl (2012) stressed the asymmetric information available to project developers and regulators that favors the former and leads to systemic overestimates of baselines.

3.2.1 Examples

The key discussions about baseline fraud relate to perverse incentives to increase emissions levels beyond historical levels in order to maximize emission credit sales revenues. Strictly speaking, this is not baseline fraud but manipulation of a project activity. This risk is seen as particularly high whenever the revenues from the sale of products are lower than those from the emissions credits. This is the case for projects to destroy HFC-23 gas, a by-product of the production of the refrigerant HCFC-22. Reference SchneiderSchneider (2011) clearly showed that such projects maximize HFC-23 production levels until the maximum level to be credited is reached, lowering them once the threshold is exceeded.

Simple fraud regarding baselines would relate to applying a baseline methodology correctly but feeding it with manipulated data. For example, the carbon content of baseline fuels could be exaggerated, especially for fuels such as coal that have a wide quality range. Fraud could include submitting samples with a particularly high carbon content to the laboratory or faking a laboratory report. There is no published example of baseline fuel emission coefficient fraud, nor has such fraud been informally discussed within the expert community. But in the context of cap-and-trade systems there is some anecdotal evidence that the sampling has been biased in order to increase free allocation or reduce surrendering requirements (Section 5.1).

The most problematic case of baseline-setting abuse is the “laundering” of “hot air” by governments. Hot air refers to emissions credits generated by applying overestimated baselines, which consider emission reductions due to the economic downturn in Russia, Ukraine, and Eastern Europe in the 1990s as if they had been reached through active mitigation policies.

Hot air laundering can even go beyond baseline fraud by inventing fictitious projects. When it became likely in early 2012 that the direct sale of hot air would be prohibited for the second commitment period of the Kyoto Protocol, Ukraine and Russia resorted to approving Track 1 JI projects with spuriously high baselines. The detailed analysis by Reference Kollmuss, Schneider and ZhezherinKollmuss and colleagues (2015) shows that Ukraine alone registered 78 projects aimed at preventing spontaneous ignition of coal piles in the Eastern Ukrainian coal basin, generating 219 million credits. Seventy-four of these were only approved once the outcomes of the UNFCCC Subsidiary Body meetings in late May 2012 suggested that direct sales of hot air would soon be prohibited (Figure 3). This project type has not been submitted to any other baseline-and-credit mechanism elsewhere since it would surely have been rejected. Baseline manipulation of these projects has two components: the assumption that 78–83 percent of the coal in the waste pile would burn, and that this would happen exactly during the years 2008–2012 (Reference Kollmuss, Schneider and ZhezherinKollmuss et al., 2015: 46). In reality, much less coal would burn, and the fires would be spread out over a very long period.

Figure 3 Approval dates and credit volumes of Ukrainian coal pile JI projects

Source: Own graph, adapted from Reference Kollmuss, Schneider and ZhezherinKollmuss and colleagues (2015: 45). The red dot indicates the only Track 2 coal pile project; all others are Track 1 projects.

Given the economic downturn in the 1990s, country-level emissions were much lower than the wrongly determined country-level baseline (all the hot air), so that Ukraine could sell a high number of credits from these Track 1 JI projects and still meet its Kyoto Protocol commitments. The example shows how politically ill-conceived country-level baselines and manipulated project-level baselines together lead to the laundering of hot air.

A less blatant, but still relevant, baseline manipulation was undertaken by Russia for projects reducing flaring of associated petroleum gas, where the baseline emissions intensity was set at almost double the level that is shown in the national emissions inventory (Reference Kollmuss, Schneider and ZhezherinKollmuss et al., 2015: 57).

3.3 Risks Related to Additionality Determination

Additionality determination has been addressed most thoroughly under the CDM. The CDM started with the approach to assess whether a project faces barriers which could be overcome due to the CDM status of the project. This approach was severely criticized by researchers (Reference SchneiderSchneider, 2009) and NGOs alike, for instance when WikiLeaks dug up a cable from U.S. diplomats in India citing a number of Indian businessmen stating that the “CDM incentive is a ‘bonus’ but not the ‘driver’ of business operations.”Footnote ⁶ As a result, CDM regulators no longer accepted a barrier argumentation but required an investment test as proposed by Reference Greiner and MichaelowaGreiner and Michaelowa (2003). Over time, the investment test became more refined, and key parameters such as threshold values for internal rates of return (IRR) were prescribed (Reference Michaelowa, Freestone and StreckMichaelowa, 2009). Reference Chen, Letmathe and SoderstromChen and colleagues (2021) showed that this substantially reduced the scope for manipulation.

After the prices of CDM credits fell significantly in 2012–2013 and project developers complained about the prohibitive costs of the investment tests, positive lists of activities automatically considered additional gained ground under the CDM. They were first applied for micro-scale activities and later also for small- and large-scale activities. While positive lists cannot be manipulated, they generate loopholes for certain activities at the margin of categories covered by the list or in sectors with dynamic technological development and reductions of investment cost (Reference Hayashi and MichaelowaHayashi and Michaelowa, 2013). For example, grid-connected solar PV installations of just below 15 MW are highly attractive commercially in many countries but still qualify under the positive list.

Reference Michaelowa, Hermwille, Obergassel and ButzengeigerMichaelowa and colleagues (2019a) discussed how to prevent manipulation of additionality determination when international carbon markets under Article 6 of the Paris Agreement allow the generation of emission credits from the introduction of policy instruments. The higher the carbon price generated by such an instrument, the higher the share of mitigation that goes beyond business as usual will be. Therefore, Reference Michaelowa, Hermwille, Obergassel and ButzengeigerMichaelowa and colleagues (2019a) proposed carbon price benchmarks for policy instruments to qualify, differentiated by country groups. For example, a carbon tax would have to exceed USD 5/tCO₂ in a developing country context to qualify for crediting.

4.1 Risks Related to Cap Stringency

Overallocation happens ex ante, when more allowances are allocated than needed, such as in the form of hot air. A surplus may accumulate during the trading phase (e.g., due to unforeseen economic recessions, reductions resulting from overlapping policies such as subsidies for renewable energy, or a large supply of cheap offsets from linked systems) but can also be the result of initial overallocation.

In practice, a surplus or an overallocation in a cap-and-trade system result in the same situation, namely that the supply of allowances exceeds the demand.

4.1.2 Examples

The risk of overallocation can be explained using three examples from the international level:

First, experiences in the Kyoto Protocol’s first commitment period from 2008 to 2012 show that due to the non-ratification of the Kyoto Protocol by the United States and the generous targets for most economies in transition, the international carbon market was overallocated from the outset. Although Annex I countries with Kyoto commitments had stated that they would refrain from buying hot air and implemented so-called Green Investment Schemes (GIS), the involvement of companies and the ability to buy ERUs from JI projects in countries in transition enabled imports of hot air into the Kyoto compliance market. This reduced the Protocol’s environmental effectiveness since for each ERU laundering hot air, one additional tonne of GHG was emitted by the buyer (Reference Kotsch, Betz, Abrell and SchwendnerKotsch et al., 2021).

Second, the overall inflow of hot air ERUs could have been manageable as the Kyoto Protocol limited the use of offsets by Annex I countries through its so-called supplementarity rule requiring that these mechanisms “be supplemental to domestic actions” (Art. 6.1 d, Art. 12.3 b, Art. 17), and stating that Kyoto units may only be used for compliance with part of Annex I countries’ commitments. However, since no agreement on the quantitative implementation of supplementarity could be reached during the Kyoto Protocol negotiations, no limitation was imposed on the use of AAUs, CERs, and ERUs in the compliance assessment of the first commitment period.

The EU did specify a limit to be applied for member states when setting the cap for the ETS sector, and EU member states seem to use fewer ERUs or CERs compared to New Zealand, as can be seen in Figure 5. The extensive use of ERUs by New Zealand (NZ) can be explained by the fact that companies regulated by the NZETS were not bound by any quantitative restrictions on the use of CERs, ERUs, and RMUs for their compliance before May 31, 2015, when the surrendering of Kyoto units was stopped. Given the high volumes of Kyoto units being surrendered for compliance by NZ entities, the country was using those units for its own compliance. In any case, it seems that NZ put in less effort domestically and allowed a high share of Kyoto units with questionable quality to be bought from other countries, thus violating the originally foreseen supplementarity rule. However, as there was no strict implementation of this rule, this had no consequences (Figure 5).

Figure 5 Use of offsets under the Kyoto Protocol’s 1st Commitment Period

Source: Own graphic based on UNFCCC compliance data. The graph shows Annex I countries’ Kyoto targets for the first commitment period and their use of the various types of Kyoto emission units for compliance.

Third, overallocation also occurred when countries which did not ratify the second commitment period (2013–2020) of the Kyoto Protocol (e.g., Japan and NZ) failed to directly cancel the remaining units in their Party and company accounts. This has led to a situation in which the demand for units from those countries has ceased but the supply of excess units still remains. The fear of lawsuits for redress and the lack of rules applicable when Parties withdraw from an agreement need to be addressed in future international carbon market regimes.

Overallocation has also taken place in regional markets such as the EU ETS, as Figure 6 shows. For almost every year in the first 10 years since the system’s introduction, the supply of allowances exceeded the demand. Only from 2014 onward, when part of the scheduled allocation was postponed to future years through so-called back-loading, the allocation stayed below the emissions. In Phases 1 and 2 of the EU ETS, the basis for cap setting were the National Allocation Plans (NAP) prepared by each member state and approved by the European Commission. This made it possible to lobby for generous allocation at the national level. Protectionist attitudes of some member states to set generous caps as well as the advantage of benefitting from financial transfers from member states with strict targets combined with a lax approval process of NAPs by the Commission in the first phase contributed to the overallocation (Reference de Sepibusde Sepibus, 2007a).

Figure 6 Overallocation and surplus in the EU ETS

Source: Own graphic based on data from the EEA EU ETS Dataviewer and from Commission communications on the total number of allowances in circulation and on international credit use in the EU ETS.Footnote ⁷ The cumulated surplus arises from the difference between all units used and the actual emissions.

In addition, Reference Betz and SatoBetz and Sato (2006) found that the enormous time pressure and some technical issues such as missing sector definitions, monitoring methodologies, and verification requirements combined with interpretation disputes with regard to coverage of certain processes created uncertainties in the data used to set the cap, which may also have contributed to the overallocation in the first phase of the EU ETS (2005–2007). Fortunately, this overallocation in the pilot phase was not bankable into the second phase (2008–2012). In the second phase, caps had to be in line with the Kyoto target of the respective member state. The Commission became stricter in assessing NAPs and did not accept cases where member states intended to assign the lion’s share of the reduction effort to the non-ETS sectors (Reference de Sepibusde Sepibus, 2007b). NAPs were also rejected if the caps were too generous, which was the case for almost all member states.

The second phase of the EU ETS illustrates how a surplus quickly builds up and is bankable into future periods. The unexpected financial crisis of 2008–2009 reduced emissions significantly (Reference Bel and JosephBel and Joseph, 2015), which, combined with higher outcomes of complementary policies such as investments in renewables, resulted in annual surpluses that accumulated over time. But there were also design elements which increased the risk of a surplus. International credits increased the allowances supply by more than 1,500 million units, which is almost the size of the surplus. As ERUs and CERs were available at substantially lower prices than the EUA price, they were used extensively up to the limit and further reduced the EUA price (see Figure 6).

The possibility of banking units from Phase 2 into Phase 3 (2013-2020) led to a situation where the surplus also affected future periods. For example, from Phase 2 to Phase 3, 1,749.5 million allowances were banked (EC, 2015).

The California ETS illustrates how overallocation can be the result of overlapping policies. In California, GHG reductions are mainly driven by Renewable Portfolio Standards to promote renewable energies as well as low-carbon fuel standards to promote investments into efficient combined-cycle gas plants. Reference Borenstein, Bushnell, Wolak and Zaragoza-WatkinsBorenstein and colleagues (2019) showed that for low reduction targets little additional abatement is needed, and therefore it is not surprising that the California Carbon Allowance prices cleared at or near the price floor and that the system is accumulating a surplus.

Linking with an overallocated system will export the overallocation to the other system. This experience can be illustrated by the link between the NZ ETS and the Kyoto mechanisms until 2015. The NZ ETS operated without a domestic cap and allowed an unlimited use of Kyoto units. With time, an oversupply of international Kyoto units such as CERs and ERUs accumulated. Their prices began to fall in 2011, which led to a decline in NZU prices (Figure 7). The sharp price decline reduced incentives for domestic reductions.

Figure 7 CER and New Zealand allowance price developments

Source: Own graphic based on data obtained from ICAP (EUA and NZ prices) and own data (CER prices). The curves show 7-day rolling averages of daily prices; CER prices are an average of Bluenext and ECX spot prices.

4.1.3 Prevention, Detection, and Enforcement

As the cap limits the tonnes of GHG emissions that can be released by the covered sectors over a certain period of time, setting the cap over a specified period at the “right” level is one of the critical decisions the regulator has to make to avoid overallocation from the outset. However, as shown in the abovementioned examples, external factors and features such as complementary policies, flexibility through banking and borrowing, linking, and rules for using offsets or for the cancelling of units in case of withdrawal may alter the cap and have an impact on the risk of overallocation or surplus. To reduce the risk of overallocation it is important to implement a solid cap setting process and specify clear rules.

Since the cap will determine the price level, politicians and regulators in the cap setting process will be exposed to intensive lobbying and will, therefore, need to be politically assertive. Achieving the right balance of scarcity in the design process is challenging and will require solid information on, e.g., historic emissions, technology, and cost development, as well as an understanding of the quantitative impact of other design features and complementary policies affecting the scarcity of allowances. For decision-makers, it is not always easy to distinguish between truly relevant information and distorted information spread by special interest groups.

Setting the cap based on emission projections is particularly prone to result in overallocation. Such a counterfactual future development is highly uncertain as it depends on economic activity and technical innovation, as well as consumer behavior. In addition, there is a bias toward inflated projections because “no government likes to predict a gloomy economic future” (Reference Grubb and FerrarioGrubb and Ferrario, 2006: 498). When the cutbacks are small relative to projections, the risk of ending up with overallocation is high. The quality of the forecast also depends on the reliability of historic data and accuracy in estimating the impact on emissions of complementary policies and measures.

Other solutions include cap-setting processes such as intensity-based caps or a flexible cap. Intensity-based caps allow the government to adjust the cap automatically to fluctuations in economic output. However, they will increase the risk that a certain emissions level is not met. They also come with some additional technical and administrative challenges and may lead to perverse incentives if they apply ex post allocation adjustments (see Section 4.3).

A flexible cap may be set in a way that it balances investor certainty with government flexibility by providing trajectories and corridors (see Figure 8). It allows for adjustment to unforeseen macro-economic developments, when new information in climate change science becomes available, when new GHG abatement technologies become available, or when the NDCs are adjusted in the process of increasing ambition.

Figure 8 Flexible cap setting on a rolling annual basis

Source: Own graphic. Periods can vary, depending on the planning horizon of the industry covered. In this graph, 10 years are the length of a period. The gateway provides the upper and lower boundaries for the company cap.

The risk of accumulating a surplus can be reduced by restrictive banking rules, as in the first phase of the EU ETS (banking only within a phase and not between phases) or by limiting the number of allowances that entities can bank, as in the California ETS. However, this would come at a cost since the flexibility of banking reduces overall compliance costs and increases economic efficiency. In addition, it reduces price volatility as demonstrated above when the end of the first and second phases of the EU ETS were compared. At the end of the first phase, the price dropped to zero whereas banking from the second into the third trading period kept the price above zero because market participants expected future scarcity.

The risk of a surplus due to the inclusion of offsets may be reduced by introducing a quantitative limit for offsets. This also needs to be properly implemented and enforced to function well, as the experience with the “supplementarity” rule on the international level has illustrated.

Another way to reduce a surplus would be to introduce a market stability mechanism (see Reference Wettestad and JevnakerWettestad and Jevnaker, 2016, Reference Wettestad and Jevnaker2019 for detailed discussions of the political processes that led to the introduction and later tightening of the EU ETS’s market stability reserve). There are two types of control mechanisms: price-based or quantity-based. The former sets a maximum and/or minimum limit on the price of allowances, the latter a minimum and/or maximum abatement requirement within a given time period. Both can be used to adjust the supply of allowances in circulation by transferring them to a reserve. Such a supply adjustment can either be temporary, meaning the allowances in the reserve are released at a later stage, or permanent, meaning the cap is adjusted and allowances in the reserve are canceled. By cancelling allowances permanently, the environmental effectiveness of the ETS is increased. For the California ETS, Reference Borenstein, Bushnell, Wolak and Zaragoza-WatkinsBorenstein and colleagues (2019) showed that there is a very high likelihood that the price will be set by the price floor or the price cap. The probability that a price between the price floor and the price cap emerges is low because the marginal abatement cost curve in California is not responsive toward higher reductions. In such settings, greater attention should be put on setting the price floor and price cap compared to setting the emissions cap, which would favor hybrid price and quantity systems. If allowances are canceled from the reserve (as in the EU ETS) there is the potential risk of a self-fulfilling prophecy: Companies invest in mitigation measures anticipating higher carbon prices in the future as a result of the introduction of the reserve. This lowers the demand for allowances and thus leads to an increased transfer of surplus allowances to the reserve. After the permanent cancellation of the allowances, allowance prices rise as scarcity increases. Hence, while the market reserve works well for preventing price decreases associated with a surplus, it might lead to stronger price rises than without the mechanism.

4.2 Risk of Overshooting the Cap

Overshooting the cap refers to a situation in which the real aggregated emissions of the regulated entities are higher than the cap set or the allowances budget. Reasons for overshooting the cap include sanctions without make-good provisions, gaps in insolvency proceedings, lax verification rules, price caps, linking with weak systems, or non-permanence when including sinks since sequestered CO₂ may be released, for example, due to a fire. Some of those reasons are interrelated and may, as mentioned earlier, be the result of lobbying for weaker regulations.

Different reasons for overshooting the cap also incur different sanctions (Reference Stranlund, Murphy and SpraggonStranlund et al., 2011):

1. 1) If a company surrenders fewer allowances than required according to its verified report, it receives a penalty for noncompliance.

2. 2) If a company underreports its emissions, the difference between its actual emissions and the required allowances may not be immediately evident. However, if underreporting is detected, sanctions are imposed for violating the reporting requirements as well as for noncompliance.

The likelihood of overshooting the cap depends not only on the level but also on the type of penalty. In the case of a fixed penalty rate, meaning a constant fine for each missing allowance, overshooting is more likely to occur the closer the penalty level is to the market allowance price. As noncompliant companies pay the fine without having to surrender the missing allowances, the result will be higher emissions compared to the cap. Such a penalty would have a similar impact as a price cap (see Section 4.1.3).

If the penalty is implemented as a make-good provision, which requires companies to make up for their permit shortfalls according to a particular ratio (e.g., a 1:2 ratio will require two allowances to be surrendered in a future period for each permit shortfall), there may be a temporary overshooting as companies are allowed to borrow allowances from future periods. The make-good cannot continue indefinitely but is usually restricted to a certain period before companies lose their operation license. Unlike the fixed penalty rate, the maximum cost of compliance under the make-good-provision is reliant on future allowance prices. Thus, greater uncertainty about future allowance prices will increase the likelihood of compliance as the level of the penalty will probably be high.

Finally, a combination of the two penalty types, a so-called mixed penalty, can apply. This would mean that there is a financial penalty, but, in addition, the missing allowances must be surrendered in the future. Thus, the overshooting risk is mainly related to fixed financial penalty rates without make-good provisions.

In case of an insolvency proceeding, it is often unclear who must surrender the allowance, which may also lead to overshooting the cap. The insolvent company may, for example, have sold emission allowances that it was allocated free of charge, and, therefore, it may no longer hold any allowances at the time of surrendering them. If such situations are not explicitly regulated by making the insolvency administrator be the one to surrender the missing allowances on behalf the insolvent company, overshooting of the cap may occur (Reference SchumacherSchumacher, 2020).

Emissions trading systems which have established a price cap will also overshoot the emission cap if the price cap is triggered and additional allowances are issued.

Linking with another cap-and-trade system will import all the risks of that other system – or even of the systems linked to it. This includes the risk of overshooting the cap, for example, in case weaker sanctions are applied in the linked systems. In addition, linking will also result in risks related to the loss of control over the design, operation, and enforcement (Reference Green, Sterner and WagnerGreen et al., 2014; Reference HaitesHaites, 2016). For example, a lower monitoring, reporting, and verification (MRV) standard or lax enforcement in the other system may reduce the effectiveness of both systems as it may increase emissions and lower prices compared to the situation before linking. In practice, however, linking of carbon markets has been much more elusive than expected, in part due to insufficient compatibility between their designs (Reference Gulbrandsen, Wettestad, Victor and UnderdalGulbrandsen et al., 2019).

Overshooting the cap will be more likely when GHG with more complex MRV or if biological sinks such as forests are included in the system. Forests bear the risk of releasing sequestered carbon due to storms, floods, fires, or deforestation. Forests also require different MRV standards and entail greater uncertainty due to more complex natural processes. In this context, the question also arises as to the definition of “forest.”Footnote ⁸

Overshooting of the cap will happen if the obligation to surrender allowances does not correspond to a one-for-one system but follows a progressive system under which one allowance is surrendered for two tonnes of CO₂.

4.3 Risk of Creating Perverse Incentives

The allocation process is particularly prone to lobbying efforts as it has distributional effects and creates winners or losers. There is, therefore, a high risk that the implemented rules serve to maximize the rent rather than being fair or efficient. There are two options: Either allowances are allocated for free to regulated entities or other beneficiaries, which means these will receive the respective value, or they are auctioned and the authority receives the value as auction revenue. The latter would be in line with the polluter-pays principle and has the advantage that it treats incumbents and newcomers equally. Therefore, auctioning should be the favored option (Reference Hepburn, Grubb, Neuhoff, Matthes and TseHepburn et al., 2006). However, free allocation is politically easier to implement because companies bear only the cost of the abatement but not of their emissions, meaning they have more money to invest in abatement opportunities. In addition, incumbents like the fact that it may create barriers for new market entrants (Reference Hepburn, Grubb, Neuhoff, Matthes and TseHepburn et al., 2006). One of the many negative effects of free allocation is the fact that companies might not realize that the freely allocated allowances have a value (the so-called opportunity cost) because they can be sold at a price on the market.

There are two major methods of free allocation, but many other combinations are possible: Grandparenting is based on a company’s historical emissions, whereas benchmarking uses a rate based on the most efficient installations or technologies within a regulated industry. Both methods may create perverse incentives for emission abatement if provisions allow for the updating of the allocation based on changes to production activity or plant capacity (Reference Sartor, Pallière and LecourtSartor et al., 2014; Reference Verde, Teixidó, Marcantonini and LabandeiraVerde et al., 2018). Further, if under benchmarking with updating based on output, the companies are not aware of the opportunity cost of holding the allowances, they may not realize cheap reduction options that would allow them to sell allowances with a net benefit. This is not efficient, as it causes market participants to forgo trades that would be more advantageous.

As allocation is often used to compensate for the additional cost of reducing emissions or buying allowances or to help ensure competitiveness between domestic and foreign companies, there is a clear trade-off between efficiency and competitiveness. The use of income from penalty payments can also create perverse incentives, especially for state-owned companies.

5.1 Monitoring, Reporting, and Verification Risks

MRV (i.e., monitoring, reporting, and verification) refers to the rules and processes that need to be in place to ensure the transparent and accurate tracking of emissions and emission reductions within the scope of a carbon market. They include the regular monitoring of the emissions at each regulated installation based on an approved monitoring plan and methodology, the submission of emissions reports, and the verification of these reports by an independent auditor.

Monitoring: First, it must be established which sources of emissions are to be monitored, including by setting a minimum size threshold for covered sources. This is usually done in a monitoring plan. As a recent case involving a chemical plant in Switzerland shows, an emission source can easily be omitted.Footnote ¹³ Second, the plan needs to include a monitoring method, such as continuous measurement and/or calculation of emissions. The method chosen will determine the reporting frequency and the verification requirements. Calculation methods are usually differentiated by tiers, where higher quality and accuracy requirements apply to larger sources and lower requirements to smaller ones (PMR and ICAP, 2021). However, there can be a trade-off between a very accurate method and the potential for abusing this methodology. For example, standardized emissions factors are not very accurate, but they are easier to verify and provide relatively limited opportunity for cheating. In contrast, if the carbon content of fuels is determined on site, which is more accurate, the sampling method can be manipulated to yield more favorable outcomes. There are anecdotal reports from the early years of the EU ETS of such samples being taken only from coal piles with a low carbon content. This is why the EU ETS now requires a sampling plan.

Reporting: Many ETSs rely on self-monitoring and self-reporting. However, to ensure the quality of the reported information, an inspection of the data is often required. Electronic systems have been established to handle the reporting and allow for easy comparison and statistical analysis (ECA, 2015). However, in the EU ETS the member states implemented their MRV systems at different jurisdictional levels. Some involve regional and/or local authorities, and there are also differences in how reporting is enforced (ECA, 2015).

Verification: Verification is often outsourced to private auditors or verifiers, which raises the question as to their trustworthiness. As a result, there is often a need to inspect the verifiers or to establish an accreditation body. In many ETSs, companies can select their own verifier. Given that verification is a recurring service to those companies, there is a conflict of interest for verifiers (see also Section 3.1). In the EU ETS, there is a lack of standards for documenting verification findings, and processes are missing on how to resolve identified problems. The probability of spot-checks by the regulator has also not been harmonized among EU member states.

5.2 Risk of Double Counting

Reference Schneider, Duan and StavinsSchneider and colleagues (2019: 180) defined double counting as “counting the same emission reduction [or removal] more than once to achieve climate mitigation targets.” Failure to prevent double counting results in greater GHG emissions than those reported by participants in the carbon market. For this reason, avoiding double counting is key to ensuring the environmental integrity of carbon markets at all levels. At the same time, avoiding double counting is necessary to maintain the trust of market participants concerning the value of the units they are acquiring.

Double counting can occur both in baseline-and-credit and cap-and-trade systems and at different stages of the market cycle:

1. – Double issuance occurs when “more than one unit [allowance or credit] is issued for the same emission or emission reduction” (Reference Schneider, Kollmuss and LazarusSchneider et al., 2015: 474). Double issuance is most likely when there is a fragmented carbon market. If there are two or more systems whose registries are not connected, it is possible that they issue units for the same reduction.

2. – Double claiming happens when “the same emission reduction is counted twice towards attaining mitigation pledges,” for example toward a country’s own climate pledge while being sold as an offset in the international carbon market (Reference Schneider, Kollmuss and LazarusSchneider et al., 2015: 475). Double claiming can also happen indirectly, for example, when a country pledges to reduce deforestation while at the same time selling credits from an efficient cookstove project (whose effect is also to reduce deforestation). Another potential source of double claiming are differences in accounting for indirect emissions, such as from electricity consumption. Baseline-and-credit systems tend to award reduction credits to those entities that undertake the emission reductions, for example through energy efficiency measures, even though the actual reduction takes place upstream during electricity generation. Double counting occurs if the reductions are counted both downstream by issuing credits and upstream by reducing the number of allowances to be surrendered.

3. – Double use means that the same allowance or credit is used more than once to attain a mitigation pledge, in either the same or different countries (Reference Schneider, Kollmuss and LazarusSchneider et al., 2015: 475). This may happen when a unit is transferred to another registry without being canceled in the original one, for example, when two ETSs are linked.

A final type of double counting occurs when emission reduction projects are used both within a carbon market and as part of climate finance provided to support developing country mitigation efforts (see, e.g., Reference Michaelowa, Espelage, ‘t Gilde and ChagasMichaelowa et al., 2020b).

5.3 Risk of Market Manipulation

Market manipulation is a form of market abuse in which market participants influence the market developments for personal gain, causing the price of allowances or credits to inflate or deflate artificially. This can be achieved through manipulative market behavior such as false or misleading transactions, price positioning, or the spread of false information. Manipulation can arise when individual market participants exploit information asymmetries even within an otherwise competitive market or by exercising market power. A dominant company which holds a significantly larger market share than its competitors has the power to restrict (extend) market output resulting in prices above (below) the efficient level, therefore generating positive real wealth without risk (Reference JarrowJarrow, 1992). Since companies participating in carbon markets typically belong to sectors with imperfect competition (e.g., the electricity sector), the exercising of market power in carbon markets represents a significant risk (Reference HintermannHintermann, 2016).

5.4 Risk of Fraud

The often-high value and electronic nature of emissions units as well as the architecture of the trading system make carbon spot markets especially prone to fraudulent activities. Factors that can affect the occurrence of fraud and losses include a system’s design features, the coherence of its legal framework, and its administrative enforceability and oversight measures. Stringent regulations that can prevent and detect fraud and effectively enforce market rules are more easily achievable in national carbon markets, where trading is governed by one jurisdiction. The risk of fraud is especially predominant where the scope of the carbon market is not congruent with its levels of governance and oversight. This is the case in (i) regional carbon markets covered by different jurisdictions, such as the EU ETS, as some governance is at the EU level (e.g., registry) while some of it is at the national level (e.g., account opening), (ii) national markets that are directly linked, as in the case of the California and Québec carbon markets, and (iii) markets that are indirectly linked through the use of international emission certificates such as CDM or JI units. Furthermore, in general, the higher the carbon price, the more attractive the carbon market is for fraudsters (Reference BussmannBussmann, 2020).

The following types of fraud can occur:

5.4.1 Value-Added Tax (VAT) Fraud

VAT fraud is a highly complex type of fraud, in which organized criminal groups divert public resources. It can occur if the amount of VAT charged on allowance transfers, or the method or the tax authority collecting it are not uniform within a carbon market (Reference Nield, Pereira and WeishaarNield and Pereira, 2016). There are two ways VAT fraudsters specializing in cross-border trades can exploit the weaknesses of heterogeneous tax regimes and gaps in market oversight: (i) missing-trader fraud and (ii) VAT carousel fraud.

In the case of missing-trader fraud, the perpetrator acquires emission allowances from another jurisdiction where they are exempted from VAT and sells them in their jurisdiction, where VAT applies, charging the domestic VAT to buyers without remitting the VAT to the tax authority. The fraudster disappears before the fraud is discovered, thereby becoming the “missing trader” (see Figure 9).

Figure 9 Missing trader VAT fraud

Source: Own graphic based on Reference Nield, Pereira and WeishaarNield and Pereira (2016).

In the more complex case of the VAT carousel, fraudsters build a network of interconnected companies located in different jurisdictions within the same carbon market, as depicted in Figure 9. To disguise the fraud, allowances are transferred along long chains of transactions through buffer companies and in a circle across borders multiple times. In each trading cycle, the importer of the allowances (i.e., the missing trader) does not return the VAT to the tax authority of the country where it is located. Moreover, each time an allowance is sold across a national border, the exporting trader (Company D), often part of the fraud network, illegitimately receives a VAT refund from the respective tax authority (Europol, 2009). This cycle can be repeated several times. Depending on the amount of VAT charged on allowances, the financial damage for the governments involved can be considerable.

A landmark case of VAT fraud occurred in the EU ETS in 2008 and 2009.Footnote ¹⁴ Emission allowances were treated as a supply of services, so that VAT was charged on the transfer of allowances. Fraudsters had an easy time in a market encompassing several jurisdictions where VAT provisions and collection are regulated at the country level, leading to a situation where fraudsters could exploit the loopholes of a patchwork of governance structures. Their actions generated losses of more than EUR 5 billion in public funds across several countries (Europol, 2009).

In April 2010, the scheme was finally stopped after a large-scale investigation conducted by the German authorities. Some 230 offices and homes, as well as a total of 150 suspects at 50 different companies, came under investigation, including German energy giant RWE AG and Deutsche Bank.Footnote ¹⁵ The missing traders in such VAT frauds are shell companies that usually exist for a few months only (Reference BorselliBorselli, 2011). An analysis of EU ETS transaction data reveals that more than half of the fraudulent volumes had been transferred by missing traders whose trading accounts were registered in Denmark, although those companies were not situated in Denmark. This suggests that the Danish registry had the loophole that enabled fraudsters to open trading accounts without the need for documentation of identity – at least during those early years of the ETS, after which such loopholes were closed in an uncoordinated way over time in the individual countries. Figure 10 shows that the fraudsters moved their operations to the Italian registry in 2010 as Italy was very late in adjusting national VAT regulations.

Figure 10 Development of VAT fraud in the EU ETS

Source: Reference WeiWei (2016).

5.5 Risk of Corruption

Corruption can affect both cap-and-trade and baseline-and-credit systems. Cap-and-trade systems are particularly prone to corruption at the stage when the regulation is being developed or when trading takes place at the government level. However, there are also specific processes where the risk of corruption is particularly high. To hide underreporting or to avoid sanctions, companies may corrupt verifiers or government officials by offering bribes.

While Reference Buen, Michaelowa and InternationalBuen and Michaelowa (2009) found that the CDM, at least in its early years, did not face significant allegations of corruption, opportunities for corruption clearly exist in various areas of baseline-and-credit systems, given that verifiers and regulators are involved in the day-to-day approval of projects and issuance of credits.

6.1 Lessons Learned from Existing Markets for Improving Governance Arrangements for Carbon Markets

Throughout this volume, we have reviewed a series of risks to the environmental integrity and economic efficiency of carbon markets. A key finding has been that those risks become larger when different markets – with diverse sets of rules, authorities, and participating entities – are connected to each other. This is in line with the findings of the literature on multilevel and polycentric governance arrangements: Decentralized authority bears the risks of insufficient, patchy, and uncoordinated regulation (Reference Bailey and MareshBailey and Maresh, 2009; Reference Biermann, Pattberg, van Asselt and ZelliBiermann et al., 2009).

How, then, can such governance arrangements for carbon markets be made more robust to loopholes, abuses, and even organized crime?

One crucial aspect is transparency. Transparency is vital for the appropriate functioning of all types of carbon markets, including the trading phase, and to reduce the risk of corruption.

In the past, sufficient, up-to-date, and reliable data on emission trends for different sectors and industries, and on the emission benchmarks of specific technologies, were crucial within individual markets in order to set a stringent cap and establish credible baselines. In particular, it was important for statistical offices to provide reliable data on past emissions and production output to allow a range of independent assessments on future trends and baseline scenarios. This should become easier in the future, where a zero target is the long-term aim, as only the emissions path from today’s emissions level to zero emissions at an appropriate “net zero date” needs to be defined. Once this path has been defined, the dispute over the correct projections becomes obsolete.

Transparency – and suitable, secure infrastructure to ensure it – is also key to avoid fraud and abuse in trading allowances. The best way to prevent or detect fraud in market transactions is to provide reliable real-time data on transfers for each account, covering not only spot but also derivate markets. Experience shows that not all regulators are willing to allow such transparency: Transparency provisions have often been overridden by arguing confidentiality concerns on cost information for new technologies or to protect buyers from liability after buying fraudulent allowances. This has led to quite different provisions on transparency for market transactions in different countries. In some countries, the disaggregated data is published only three to five years after the fact, which makes it impossible for experts to monitor what is happening in the market. If data are not timely and publicly available, market activity cannot be monitored sufficiently. A lack of transparency leads to a creeping erosion of the market’s stringency because problems and abuses are not discovered and addressed.

Preventing fraud such as double-issuance and double-use of carbon market units requires robust registries and transaction logs, including monitoring of transactions between exchanges and across different markets. When linking markets, therefore, provisions need to be made for sharing information between the regulators of the participating markets in real time, ideally by creating a common registry. Registries need to be secure to prevent hacking and theft, for example, by ensuring registry account security through KYC checks.

The problems caused by a lack of transparency may be compounded by insufficient staffing of regulators. In the EU ETS, for example, the staff of national-level regulators (for example, the German Emissions Trading Authority [DEHSt] with around 160 employeesFootnote ²⁰) is much larger than the staff in the offices of the overall regulator in Brussels (the section on European and International Carbon Markets at the Directorate-General for Climate Action, with 68 staff membersFootnote ²¹). This is again a problem of multilevel governance, with national-level units often having larger budgets than those at the supranational level. In the EU ETS, many administrative tasks such as the registry and allocation of allowances have been moved from the national to the European level, but the number of staff seems not to reflect those changes.

If more staffing is allocated to the higher market level rather than to individual country regulators, then policy, monitoring, and enforcement can be made more consistent. In addition, insufficient staff means that the regulator itself has limited capacities to monitor market transactions, detect irregularities, and act on them. At the opposite extreme, overstaffing can also become a problem, as the experience with the CDM staff at the UNFCCC Secretariat has shown. There, the unexpected level of resources from the administration fee levied on project registration and credit issuance led to an increase in staffing and influence exerted by the Secretariat on regulatory decisions (Reference Michaelowa and MichaelowaMichaelowa and Michaelowa, 2017). At some point, this may become inappropriate. The resulting governance issue is, therefore, to find the right staff size – and to allocate a sufficiently large budget to finance it – without falling victim to Parkinson’s law of growing bureaucracies.

A third, crucial, governance aspect, particularly in supranational and interlinked markets, is the compatibility of legislation, regulations, and even industrial structures, and the handling of unforeseen regulatory loopholes due to incompatibilities or changes in one of the systems. When loopholes exist and compatibility is lacking, the risk of abuse or fraud is correspondingly higher.

A fourth governance aspect is related to the public law principle of legality. It requires for market rules and provisions to be properly specified to be enforceable. One example is the case of supplementarity of the use of offsets, which was recognized as an important principle to avoid watering down emission targets under the Kyoto Protocol. Even though supplementarity is mentioned in the Kyoto Protocol in three places, and although it was discussed at length in the subsequent negotiations, an acceptable or allowed level for a party’s use of Kyoto units was never agreed. The practical result is that it became impossible to enforce supplementarity, at least internationally, although this is a principle under the Kyoto Protocol.

Another example from the EU ETS is that there is a lack of information on how often spot checks on reported emissions are undertaken by authorities. Often, the exact fines issued for violations in case of detection are also unknown, which makes it impossible for companies to estimate the potential penalty. In addition, authorities believe that the cost of spot-checking small companies is likely to exceed the potential revenue from fines, which may encourage further criminal behavior from small companies.

Regulatory roles in carbon markets are frequently assumed by private actors. While this approach lightens the load of overburdened public officials, and while competition may increase the effectiveness of these private regulators, it may also lead to conflict-of-interest situations. In the EU ETS, for example, external advisers were not only involved in establishing the regulations, but they later also acted as verifiers of monitoring reports of participating installations. As a result, there is an incentive for these advisers to ensure that the rules they help design are more complicated than necessary as this may make their involvement in later assignments more extensive and, therefore, more profitable. Such situations have been observed in the Australian ETS process, where law firms were involved in writing the ETS bill, or in the EU ETS, where consultants drafted monitoring and verification guidelines.

Ideally, therefore, the consultants involved in establishing market rules should be different from those executing them. However, in a new market with relatively few experts, this may be difficult to achieve. This is why it is important for the authority to ensure the quality and integrity of verifiers through accreditation processes and spot checks by the authorities. Finally, to ensure an independent judgment, verification should be paid for by the authority rather than the regulated entity, and a rotation system for verification assignments should be introduced.

A final consideration relates to dealing with unexpected circumstances and crises. Frequently, regulators react to a crisis by relaxing certain rules to help the market cope with the situation. An example of this are the new baseline rules under CORSIA to deal with the effects of the coronavirus crisis on the aviation sector. When such requirements are relaxed, there is a risk that they become sticky and remain in place if there is no clear plan for going back to the stricter regulations once the crisis is overcome. This might reduce environmental effectiveness in the long term. Measures in reaction to a crisis should, therefore, always include provisions dealing with the end of that crisis.

In the next sections, we provide an overview of future challenges of carbon markets. We start with a medium-term perspective focusing on the international negotiations around Article 6 of the Paris Agreement and on voluntary carbon markets. Article 6 is still highly contested, and no agreement on detailed rules has been reached at the time of writing. International rules may become even more important in the long term when countries need to achieve the net-zero targets required by the Paris Agreement. Since suitable storage capacities or other prerequisites for successful implementation of negative emissions technologies vary substantially between countries, carbon markets are likely to continue playing a role in the long term. In the final section, we present our thoughts on likely amendments to and open research questions about carbon markets in a net-zero world.

6.2 Challenges for International Carbon Markets under the Paris Agreement

Article 6 of the Paris Agreement refers to two types of international markets. Article 6.2 allows countries to cooperate without any international oversight other than principles and certain reporting requirements. This is akin to the first track of JI and likely to generate similar problems as discussed in Section 3. Given that there are no limitations to the type of activities, collaboration under Article 6.2 may include transboundary or linked cap-and-trade systems as well as baseline-and-credit systems. By contrast, the Article 6.4 Mechanism is subject to international regulation through a supervisory body (SB). It is a baseline-and-credit system building on experiences from the CDM and the second track of JI. Both markets will generate internationally transferred mitigation outcomes (ITMOs).

According to our own analysis, out of 88 countries that had submitted updated NDCs to the UNFCCC by end of July 2021, 68 countries, or 77 percent, have explicitly stated that they want to use Article 6. Five of those have declared they only want to buy ITMOs, while another five are open to both buying and selling. Forty-two countries have clearly stated to aim at selling only, while 16 do not provide a clear statement. So, currently there is a clear imbalance between supply and demand.

Our discussion needs to be seen against the backdrop of the detailed rules for Article 6, which were only decided at COP26 in late 2021. Many details will need to be worked out over the next few years, as experience with the CDM suggests. There are, therefore, many opportunities to take into account the key lessons learned from past markets.

The intense calls to prevent double-claiming (see Section 5.2) led to a robust accounting system with corresponding adjustments without exceptions to ensure that any ITMOs sold to other countries or entities are not counted toward the host country’s own target. As long as the NDC target is more stringent than business-as-usual emissions, this is a good safeguard to prevent Article 6.2 generating ITMOs with low environmental integrity. Problems arise if the NDC target is not binding, which would generate hot air. As shown in Section 3, absence of international oversight here is likely to lead to manipulation of international carbon markets. This means that the detailed reporting requirements, which are still under negotiation in 2022, need to be as detailed and transparent as possible to enable NGOs and other stakeholders to recognize signs of manipulation and raise the alarm (Reference Michaelowa, Espelage, ‘t Gilde and ChagasMichaelowa et al., 2020b). Moreover, UNFCCC expert review teams should be given a strong mandate to uncover manipulation in monitoring reports and to request countries to adjust their programs. However, under Article 6.2, the key reporting will take place on an aggregated, jurisdictional level rather than in terms of individual activities. In the CDM, aggregated reporting became much less transparent than originally envisaged due to strong lobbying of credit buyers based on competitiveness concerns, which led various jurisdictions to withhold disaggregated data on transactions of CDM credits (Reference Michaelowa, Censkowsky and EspelageMichaelowa et al., 2021a). The reporting and transparency rules for the Article 6.2 mechanism therefore need to be as disaggregated as possible to enable identification and elimination of problematic credits.

The second-best approach to ensuring that Article 6.2 does not lead to manipulation would be the formation of clubs of seller and buyer countries which would agree to adhere to minimum standards. The first club of this type was a group of 32 countries underwriting the “San José Principles for High Ambition and Integrity in International Carbon Markets” orchestrated by Switzerland and Costa Rica in December 2019.Footnote ²² So far, however, the operationalization of these principles remains unclear. For example, the Swiss KliK foundation, in its public calls for Article 6 pilot activity submissions, applies relatively broad criteria, which are susceptible to manipulation and abuse and certainly no better than the criteria for validation under the CDM. In contrast to that, Sweden has hired the Gold Standard to certify all Article 6.2 ITMOs. Given that the Gold Standard has built its reputation on generating high integrity credits, this safeguard mechanism is likely to work.

The principles to ensure environmental integrity under the Article 6.4 Mechanism are quite strong. For the first time, the additionality definition makes sense, requiring to show that “the activity would not have occurred in the absence of the incentives from the mechanism.” All relevant national policies need to be considered and locking in “levels of emissions, technologies or carbon-intensive practices” is to be prevented. Baselines shall be set below business-as-usual, and need to be aligned to the long-term targets of the Paris Agreement. This allows application of the concept of an “ambition coefficient” (Reference Michaelowa, Michaelowa, Hermwille and EspelageMichaelowa et al., 2021b; see also Section 6.3).

In operationalizing these principles in the next years, the rich lessons learned while developing baseline and monitoring methodologies need to be incorporated. Given that developing new Article 6 methodologies from scratch would take many years and be costly (Reference Michaelowa, Brescia and WohlgemuthMichaelowa et al., 2020a estimated that the development of one new methodology requires USD 0.1–0.2 million), an International Initiative for Development of Article 6 Methodology Tools (II-AMT) was launched in 2022, with a group of international methodology experts from all continents to generate tools that can be added to Kyoto mechanism baseline and monitoring methodologies in a modular fashion (Perspectives, 2022). Such Article 6 tools will contain clear definitions of key concepts serving as guidance on what to consider for alignment with host countries’ climate and sustainable development priorities. Guidance will also be included on how to consider existing and planned policies in additionality determination and baseline setting as well as on different concepts related to “conditionality” in NDC targets. The tools will support the alignment of Article 6 activities with the NDC implementation periods and ensure an update of key parameters in reference scenarios and crediting baselines in line with both these implementation periods and the pace of technological change and innovation in different sectors.

The tools will also facilitate reporting by the host country on how its engagement in Article 6 cooperation contributes to NDC implementation, prevents new burdens such as those generated by selling off “low-hanging fruit,” and fosters sustainable development. They aim at redefining and reconceptualizing the concept of additionality to cover three “shades” of additionality: financial additionality (going beyond a commercially viable business-as-usual project), regulatory additionality (not being mandated by law or regulation), and target additionality (not being required for host country achievement of (unconditional) NDC targets). While striving for high integrity, the tools aim to be practical in nature and be no undue burden for project developers. Their development will be accompanied by a thorough assessment of their applicability in different country contexts and of related transaction costs.

Host country stakeholders will play a key role in governing both Articles 6.4 and 6.2. This entails risks regarding the manipulation of baseline and additionality determination as well as corruption in the context of authorization. Thus, building up the capacity of host country institutions, consultants, and activity developers to deal with issues related to the Article 6 mechanisms and addressing potential conflicts of interest through appropriate mechanism design (e.g., transparent tenders or rewards for discovering abuse) is crucial to prevent loss of environmental integrity. Moreover, seller country governments need to be enabled to consider the opportunity costs of selling ITMOs due to corresponding adjustments when assessing the benefits obtained from Article 6 cooperation. Such an understanding improves policymakers’ judgment of the appropriateness of prices paid for credits, which reduces corruption risks.

Another crucial issue is the link between Article 6 and voluntary carbon markets. While any ITMO will be subject to corresponding adjustments, non-authorized Article 6.4 emission reductions will not. Currently, an intense debate is ongoing among voluntary carbon market stakeholders regarding the need for corresponding adjustments for internationally transferred credits. This now leads to two potential approaches for voluntary market stakeholders as shown in Figure 11 whereas the potential uses of credits from the voluntary carbon market depend on whether they implement corresponding adjustments or not.

Figure 11 The two possible approaches for the voluntary market under Article 6

Source: Reference MichaelowaMichaelowa (2021). CA = Corresponding adjustment; OMGE = Overall Mitigation in Global Emissions; GS = Gold Standard; VCS = Voluntary Carbon Standard; JCM = Joint Crediting Mechanism. Any use of ITMOs requires corresponding adjustments. But voluntary market credits may continue to be transferred outside the Article 6 infrastructure.

According to the International Carbon Reduction and Offset Alliance (ICROA), corresponding adjustments are unnecessary because “carbon reductions financed by the [voluntary carbon market] are not exported from the host country” and “voluntary activity does not lead to double counting at the UN level because carbon reductions are recorded only once by the country hosting the mitigation activity” (ICROA, 2020: 1), and not by the country where the buyer company resides. The largest voluntary carbon market standard provider Verra supports a similar view.

A company buying voluntary credits will, of course, use them to offset emissions in its home country to prevent becoming subject to a mandatory mitigation policy. If a mitigation outcome is counted toward a host country NDC target as well as claimed to offset emissions caused elsewhere, this is a clear case of double-claiming. This causes the argument of ICROA and Verra to crumble, which is acknowledged by the Reference StandardGold Standard (2021). It should be stressed that voluntary carbon markets can contribute to enhancing global mitigation only when they incentivize mitigation beyond host country commitments and contribute to closing the “ambition gap” that characterizes most current NDCs. If voluntary carbon markets are not adjusted for accordingly, there is a high risk that host countries will fail to implement mitigation policy instruments to the extent that would otherwise be necessary to achieve the NDC targets.

6.3 Open Research Questions: How to Make Carbon Markets Consistent with a Net-Zero World

In addition to the still not fully resolved issues concerning Article 6 and voluntary carbon markets, there are many open questions as to how carbon markets can be made compatible with a net-zero world. Article 4 of the Paris Agreement states “that a balance between anthropogenic emissions by sources and removals by sinks of greenhouse gases in the second half of this century” is necessary “in order to achieve the long-term temperature goal.” To achieve net-zero targets or GHG neutrality by 2050, the IPCC (2018) and the IEA (2021) refer to substantial investments in carbon capture and storage (CCS) and carbon dioxide removal (CDR) to offset the difficult-to-reduce GHGs from agriculture, industrial processes, and aviation. The IPCC also foresees CDR for emission overshoots that would otherwise take the planet beyond the carbon budgets consistent with a 1.5°C temperature target.

Baseline-and-credit systems face specific challenges in a net-zero world: What, for example, is an appropriate approach to baseline setting for emissions reductions? Ensuring that the baseline is in line with achieving the long-term temperature goal of the Paris Agreement and with host countries’ long-term Low Emissions and Development Strategies or carbon neutrality commitments requires bold new approaches. Baselines defined in intensity terms (volume of GHGs per unit of production of a good or service), as has been the norm under existing baseline-and-credit mechanisms, create the problem that even if the emission intensity falls substantially, and credits are allocated for such a reduction, overall emissions fall much less or may even rise because production levels increase.

So far, it has been argued that such an approach is important to address the “suppressed demand” for goods and services in poor countries. For this reason, the application of an emission intensity approach needs to be allowed to continue. It must, however, be combined with an “ambition coefficient” to ensure that, at the point in time where emissions are to reach net zero, any baseline for emission reduction activities also reaches zero. The most straightforward approach for such an ambition coefficient is to define an emissions path from today’s emissions level to zero emissions at an appropriate “net-zero date” for each host country of a baseline-and-credit mechanism (Reference Michaelowa, Michaelowa, Hermwille and EspelageMichaelowa et al., 2021b). Applying the principle of common, but differentiated, responsibilities means that the net-zero date for industrialized countries would be quite soon, probably in the 2030s. For least developed countries (LDCs), the net-zero date would of course be much later, perhaps around 2070. At each point in time, the ambition coefficient would be equal to the percentage of today’s emissions reached for that specific year on the downward sloping path, reaching zero in the net-zero year. This approach is depicted in Figure 12.

Figure 12 Application of an ambition coefficient toward a net-zero world

Source: Own graphic based on Reference Michaelowa, Michaelowa, Hermwille and EspelageMichaelowa and colleagues (2021b). The ambition coefficient would be differentiated according to level of development.

In the context of Article 6.2, buyer clubs could decide on the appropriate level of ambition coefficient, while for Article 6.4, this decision could be taken by its Supervisory Board. Ambition coefficients cannot be manipulated openly; obviously, the process of setting the net-zero year is vulnerable to corruption. As suggested above, a transparent approach would reduce this risk.

Dealing with countries whose emissions deviate increasingly from the ambition-compatible path would be challenging as emissions continue, while related emission reductions cannot create any more credits. This generates a strong incentive problem. Naming and shaming them will probably not be sufficient to bring these countries on track.

The situation is different if we assess what changes are necessary to cap-and-trade systems. To understand what is traded in a net-zero world, we first look at the technological and biological options to achieve negative emissions. Figure 13 classifies the removal options and compares them with today’s common emission reduction approaches, showing how they have been covered by carbon markets so far. This allows us to understand the types of incentive that already exist and what types still need to be developed. Cap-and-trade systems give industrial plants or fossil fuel power plants an incentive to invest in CCS as they allow companies to buy fewer CO₂ allowances and even to sell allowances if they have received free allocation. However, this would lead only to negative emissions if the burned fuel is biogenic; otherwise, it will only lead to zero emissions.

Figure 13 Carbon removal options in a net-zero world

Source: Own graphic. Green relates to negative emissions, blue to zero or near-zero emissions, black to reduction compared to a baseline situation, and red indicates that there may be temporary sequestration but that at the end of the usage process the carbon is emitted into the atmosphere. BiCRS: Biomass Carbon Removal and Storage.

As natural CO₂ sinks due to afforestation and reforestation or sequestration in the soil have usually not been covered by cap-and-trade systems – except in New Zealand – they may be included in offsetting systems. Such CDR options involve taking CO₂ from the atmosphere and binding it in charcoal (biochar) or storing it in vegetation or in the soil. These natural sinks are relatively inexpensive, and there is ample experience in dealing with them. However, such storage is short-lived, and there is a high risk of reversal when forest fires or other events release the CO₂ back into the atmosphere. This risk needs to be considered when designing an offset market with natural carbon removal.

Baseline-and-credit systems may also provide incentives for investments in removal technologies that allow CO₂ to be captured directly from the atmosphere and stored in suitable facilities underground. This process is called direct air-carbon capture and storage (DACCS). In order to achieve permanence, the captured CO₂ can be dissolved in water which is pumped into basaltic rock. In about two years the CO₂ mineralizes in pores of the basalt and is therefore stored very safely. Negative emissions can also be achieved by using biomass for energy production – green methane – and subsequently storing the CO₂ emissions in a process called “bioenergy with CCS” (BECCS). A risk of BECCS that would need to be addressed by the offset market is the probability of increased emissions when biomass production leads to additional forest clearance.

Insurance and buffers are options for addressing the risks of reversal and leakage under baseline-and-credit systems (see Section 4.2.3).

For the removal of GHGs, the baseline methodologies could continue to apply the current approaches until countries take up “net-negative” targets that involve achieving a specific amount of GHG removal. Once negative targets are introduced, a new approach could be for removal activities to only become eligible for ITMO transfers once the domestic removal target has been achieved. That would, however, lead to strong swings in eligibility over time. An alternative would be to define a negative baseline at a removal intensity consistent with achieving the overall removal target within the relevant NDC period. This approach would generate a better incentive to continue to engage in the market. If removal takes a larger or even a monopoly role in international carbon markets, issues related to fraud regarding storage site MRV will become highly relevant.

As illustrated in Figure 13, most of the CDR technologies require CCS. As not all countries have suitable storage sites, and residents frequently react with resistance if CO₂ is to be stored on their doorstep, it is likely that storage will result in some international transfers with the captured CO₂ being stored offshore and in former gas and oil deposits, locations where the required infrastructure and technical knowledge are also already available. However, the transfer of CO₂ to distant locations results in long transport routes, and CO₂ leakage may well be a risk along the way, which would need to be addressed.

If capture, transport, and storage are covered by a cap-and-trade system, allowances covering any leakage would need to be surrendered, as is done in the EU ETS with CCS and in the New Zealand system with leakage or reversal in afforestation and reforestation.

Direct air capturing with usage, harvested wood products, as well as carbon capture and usage (CCU) are further technologies that could help, at least temporarily, to shift emissions to the future. However, these options cannot reduce, avoid, or remove emissions permanently. Therefore, a different accounting method needs to be applied.

If carbon markets integrate CCS and removal units, the market will determine what technology will be deployed. As illustrated in Figure 14, as soon as classic abatement options including green hydrogen become more expensive than buying removal units from DACCS and BECCS (which function as backstop technologies), they will be replaced. The cheapest technologies will be used first. As a result, a market price will emerge. This must be high enough to finance negative emission technologies.

Figure 14 A “net-zero” cap-and-trade system

Source: Own graphic.

Given that CDR will require high carbon prices and that free allowance allocation is not possible in a net-zero world, carbon border adjustment mechanisms (CBAM) can help to reduce the impact on competition. Carbon markets could be integrated into such mechanisms; for example, an exemption from paying the CBAM rate by surrendering a removal or reduction credit could be granted.

There are a number of open questions with regard to carbon markets under a net-zero target:

How are the current blueprints of carbon markets compatible with CDR as biomass plants, for instance, are usually not covered by cap-and-trade systems? Will there be a separate market for emissions removal units or will markets be integrated, and will they apply qualitative and quantitative restrictions? What will be the share of CDR and emissions reductions in achieving the net-zero targets if there are separate markets? Who should be awarded the removal unit, and how will the value be divided between capturer, transportation, and storage provider? How can co-risks such as land conflicts, food security, loss of biodiversity and co-benefits such as enhanced biodiversity be addressed and reflected adequately?

These are all questions that will need to be answered by future research on carbon markets.

6.4 Concluding Remarks

At this point, the reader may be somewhat overwhelmed by the complexity of designing carbon markets and all of the risks involved. Given that this volume focuses explicitly on such risks, this is not unexpected. One may be tempted to ask oneself, however, whether it is at all possible to design functioning carbon markets. Given that they are being adopted by a growing number of jurisdictions, however, carbon markets are probably here to stay. If designed well, they can be an important tool for closing the gap between the Paris Agreement’s global mitigation goals and countries’ NDCs.

As this volume has shown, carbon markets have given rise to regulatory challenges since their inception, yet the policy responses to these challenges have also matured and grown in sophistication over time. Studying the complex regulatory frameworks that have developed in existing cap-and-trade and baseline-and-credit systems can offer valuable lessons for the regulation of emerging and future carbon markets.

Regulatory challenges in these markets will continue to evolve, however, as the climate policy objectives they serve become more ambitious and converge around net-zero emissions, the value of traded units gradually increases, and markets become more interconnected, whether through bilateral linkage or multilateral cooperation frameworks such as Article 6 of the Paris Agreement. New digital tools, such as software-based analysis of market data and remote sensing for compliance control, may help manage these challenges as they become available and more widely used. Ultimately, however, the precise nature of such future challenges is difficult to predict.

In such a context of uncertainty, future decision-makers can derive guidance from the general principles of carbon market regulation set out at the beginning of this volume. At any rate, if carbon markets are to play a role in advancing the global effort to decarbonize our economies, sound regulation will only gain – not lose – in importance going forward.