Why carbon leakage matters and what can be done against it

Abstract

Global greenhouse gas emissions continue to climb, meaning increasingly stringent climate policies are needed if we are to meet the Paris Agreement targets. However, policies that curtail emissions in one region might increase emissions in another through the transfer of emission-intensive production across borders. Such spillover of emissions, mostly via international trade, can cause what is referred to as “carbon leakage.” This primer first provides an overview of the underlying mechanisms, empirical evaluation, and future projections of carbon leakage. It then illustrates potential anti-leakage policies, including free allocation of emission permits and border carbon adjustment. Well-designed anti-leakage policies that create economic and environmental benefits and ensure fairness and justice are clearly required. The 2050 net-zero commitments announced by over 100 nations offers an opportunity to form a climate club with common border measures to prevent carbon leakage while incentivizing ambitious climate actions. Global greenhouse gas emissions continue to climb, meaning increasingly stringent climate policies are needed if we are to meet the Paris Agreement targets. However, policies that curtail emissions in one region might increase emissions in another through the transfer of emission-intensive production across borders. Such spillover of emissions, mostly via international trade, can cause what is referred to as “carbon leakage.” This primer first provides an overview of the underlying mechanisms, empirical evaluation, and future projections of carbon leakage. It then illustrates potential anti-leakage policies, including free allocation of emission permits and border carbon adjustment. Well-designed anti-leakage policies that create economic and environmental benefits and ensure fairness and justice are clearly required. The 2050 net-zero commitments announced by over 100 nations offers an opportunity to form a climate club with common border measures to prevent carbon leakage while incentivizing ambitious climate actions. At the time of writing, 127 countries representing more than 60% of global emissions, about two-thirds of global GDP and more than half of the global population have adopted “net-zero” targets to phase out activities linked to the generation of greenhouse gas (GHG) emissions, including key players, such as China, the US, EU, and Japan. Even though actual measures for their implementation are still under deliberation, the rapid spread of these targets might well be considered a promising prospect for climate policy that is in line with the targets of the Paris Agreement. However, different levels of ambition in different world regions could turn into a serious obstacle for effective climate policy. A frequent concern in this regard is that stringent climate measures in some countries could be undermined by increasing emissions elsewhere—frequently labeled “carbon leakage.” Without appropriate countermeasures, carbon leakage can undermine the effectiveness of mitigation activities and decrease public support for climate policy. This primer explains the underlying reasons for carbon leakage and discusses empirical evaluation of leakage as well as projection for different future scenarios. It then provides an overview of possible policies to address leakage and illustrates their legal and political implications. The academic literature identifies three channels due to which more ambitious climate policies in one part of the world may increase emissions elsewhere, thus undermining the effectiveness of climate measures to at least some extent (see Figure 1). First, climate measures in one world region lower the demand for fossil fuels and hence reduce their world market price. As a reaction to this price decline, other parts of the world can be expected to increase their consumption of fossil fuels, which offsets at least some of the achieved emission reductions (this mechanism is often referred to as the “energy market channel,” as shown in Figure 1). Furthermore, the mere anticipation of stringent climate policies could provide an incentive for fossil fuel owners to accelerate their extraction, which would result in so-called intertemporal leakage. Second, climate policy likely increases energy prices in implementing regions. This might shift patterns of trade specialization toward increasingly carbon-intensive imports and provide incentives for energy-intensive industries, such as steel, cement, and chemicals, to migrate to regions with laxer climate policies (the “competition channel”). Such leakage could entail the risk that climate measures harm the competitiveness of domestic energy-intensive industries without achieving actual reductions in global emission. Third, as emission reductions in one part of the world also benefit other regions, climate change mitigation constitutes a global public good. It hence faces problems similar to those arising for protecting the ozone layer, safeguarding biodiversity, preventing pandemics, or maintaining international security: in all these settings the private benefit of contributing to a global public goods are smaller than the collective benefits. Hence, with individual countries weighing their own interests higher than those of others, such public goods will be underprovided. From a game-theoretic perspective, climate policies in different world regions can thus be considered “strategic substitutes,” such that for each country the best response to emission reductions made by other countries is to allow itself to increase its own emissions (the “free-riding channel”). This problem might become even more serious with the possibility to use “offsets” to pay other countries instead of reducing domestic emissions, as such payments could incentivize some countries to lower their climate change mitigation efforts in order to receive higher payments in the future. Different approaches have been used to empirically assess carbon leakage. One approach examines the impact of different energy prices across regions on trade flows and relates these estimates to energy price changes that might result from climate policy. Another empirical strategy consists in comparing similar firms that are covered by or excluded from climate measures (such as the EU Emission Trading Scheme [EU ETS] or the UK climate levy). Both approaches consistently yield very similar results, finding limited evidence of carbon leakage in only very few energy-intensive industries. A probable explanation for this finding is that, for the large majority of industries, energy costs play a rather minor role, very rarely accounting for more than 5% of total production costs. Hence, factors, such as proximity to key markets, access to skilled labor, as well as a favorable political and institutional environment to conduct business, might be more important than energy prices. Yet, this should not be regarded as evidence that carbon leakage is unproblematic. Until now the costs of climate change mitigation have been rather moderate due to the relatively low ambition of abatement targets. In addition, firms have frequently been shielded by some form of anti-leakage measures, such as the free allocation of emission permits under the EU ETS, which will be discussed in more detail in the section “anti-leakage policies.” More stringent climate targets in the future are likely to be accompanied by tighter climate policies, which might substantially increase carbon leakage, especially if they are only adopted by a subset of countries. The debate about carbon leakage is closely related to the question of who should be responsible for emissions released in one country to produce goods and services that are exported and consumed in other countries. Under the traditional production-based accounting (PBA) of emissions, which is used in, e.g., the United Framework Convention on Climate Change, each country is assigned those emissions that are released within its national territory. This approach is clearly unsuitable to capture the complex interrelationships between consumption pattern in one country and production decisions (and thus GHG emissions) in other countries. As an alternative, consumption-based accounting (CBA) of emissions has been proposed, under which countries are assigned all emissions generated to meet their final consumption. CBA accounts are computed by adding all import-related emissions to and subtracting all export related from a country's PBA account. The difference between import- and export-related emissions is frequently considered as a net import of emissions and labeled an “emission transfer.” Studies based on input-output models have demonstrated sizable emission transfers mainly from low-income to high-income countries. Some have argued that these emission transfers are an indication that the latter are shirking their responsibility to reduce global emissions by maintaining their environmentally harmful patterns of consumptions while outsourcing carbon-intensive industries to other countries. However, more recent findings have clearly shown that this CBA approach would incorrectly identify carbon leakage that has never been generated. The main reason for this is that this way of accounting for emissions does not credit countries for cleaning up their export sectors. For instance, when some countries lower the carbon content of their exports, while trade patterns and production structures in other countries remain unchanged (i.e., the emissions of their imports remains constant), the CBA approach (i.e., import emissions minus export emissions) will show an increase of net import emissions, which will be taken in to indicate carbon leakage even though no leakage (i.e., through the relocation of production from high-income to low-income countries) has ever occurred. In contrast, the lower export-related emissions would decrease the amount of emissions for the import countries, and it could even reduce the overall emissions globally. This has beneficial impacts on the climate, which, however, are not reflected in the CBA account of the country decarbonizing its exports. Consequently, a desirable property to assess individual countries' responsibility for global GHG emissions is that changes in the employed emission accounting scheme should display a one-to-one relationship with changes in global emissions. In this way, any change in production or consumption patterns in a certain country affecting global emission would be directly reflected by an identical adjustment of its emission accounting. PBA and CBA both fail to meet this requirement: whereas PBA fails to consider emissions generated abroad to meet domestic demand, CBA does not take into account emissions generated domestically for foreign demand. Thus, they do not appear as useful tools to fairly assess carbon leakage, and there is no reason to assume that one scheme is inherently better suited than the other to track the impact of developments in one region on the global climate. Recently, a more appropriate accounting scheme has been proposed. This approach accounts for export-related emissions on the basis of the global average of the respective sector's emission intensity. As CBA is calculated by adding (subtracting) emissions related to imports (exports) to the PBA, exports of, say, machinery produced free of carbon would not change a country's CBA inventory relative to PBA. By contrast, under “technology-adjusted CBA,” machinery exporters can subtract the emissions that correspond to the global average carbon intensity of producing machinery from their emission inventory. This accounts for the fact that their exports have avoided emissions that would otherwise have occurred for machinery production elsewhere. On this basis, a more complex picture of responsibility for trade-related emissions emerges. For instance, the US is indeed found to contribute to emissions elsewhere, as trade patterns have shifted in a way in which the US increasingly imports carbon-intensive products. However, EU exports of relatively energy-intensive goods and services are now produced with a lower carbon intensity, which avoid the higher emissions that would have been generated had the goods been produced outside of the EU and exported to the EU. That is, the narrative of high-income countries consistently outsourcing their emissions to other countries cannot be upheld under this novel perspective. Accounting schemes that acknowledge the joint responsibility of producers and consumers and divide trade-related emissions between them, for example, in proportion to the economic benefits derived, are an alternative and arguably fairer route to emission accounting in an increasingly interconnected global economy. Numerous studies assess the potential for carbon leakage based on numerical computable general equilibrium models that include a stylized representation of global energy markets, production, and consumption of goods and services, as well as international trade. Estimates of leakage rates (i.e., the share of emission reductions in one world region compensated by increases in other parts of the world) provided in the literature show a broad range. On average, models project leakage rates of about 14%, with the large majority of studies falling in the range of 5%–20%. Leakage rates depend on the differences in mitigation costs across regions, which are determined by the global emissions reduction target as well as the distribution of emission reduction efforts across regions. In addition, the modeled economic mechanisms and policy instruments can also affect the estimated leakage rates significantly. Some studies show that market structures characterized by oligopolistic competition (i.e., market prices determined by only a few large firms) and elastic demand for fossil fuels (i.e., small price changes resulting in large changes in demand) generate relatively high rates of carbon leakage. That is, firms in an oligopoly market and in unregulated regions without emissions regulations could strategically respond to output reductions elsewhere (resulting from climate policy) by raising their output, and eventually increase the carbon leakage. Furthermore, lower fossil fuel demand in regions with climate policy is projected to substantially depress world market prices for fossil fuels and raise consumption in other regions. Others, however, point to the potential of “negative leakage,” i.e., emissions reductions in one world region resulting in lower emissions elsewhere. This could occur by, for instance, decreasing costs for clean technologies due to technological learning, which accelerate their deployment. Another possible mechanism runs through the decreased demand for natural gas in countries with ambitious emissions reduction targets. This could increase gas use in other countries and push coal, which has a roughly two times higher emission intensity, out of the global energy system. Hence, even though carbon leakage seems unlikely to completely undermine climate policies adopted by front-runners, it might reduce their effectiveness. For this reason, dedicated policies to reduce carbon leakage seem necessary. Several policies to deal with carbon leakage have been proposed. To date, the most widely used anti-leakage policy is the free allocation of emission allowances in the context of tradable permit systems, most importantly the EU ETS. Typically, in such schemes firms receive emission permits in proportion to their output, using a reference (such as the 10% most-efficient firms in the respective sector) as a benchmark. As firms receive an amount of emission permits that is proportional to their output, but independent of its carbon content, they have an incentive to employ carbon-saving technologies—more carbon-efficient production methods would yield the same amount of output with less emissions, and therefore lead to emission permits surplus, and extra revenues can be generated by selling the excess permits to other firms in the market for emission allowances (in the case that the amount of freely allocated permits is below actual emissions, firms that employ carbon-saving technologies could reduce the requirement to purchase additional permits). At the same time, the free allocation mechanism under the EU ETS reduces the incentive to substitute domestic production with imports, as each unit of output produced domestically will be rewarded with a certain amount of emission permits. This is desirable in terms of leakage protection, but problematic from a climate perspective, as it de facto provides a subsidy for the production of carbon-intensive goods. Additional consumption charges on the use of energy-intensive goods (independent of the respective carbon content) are one avenue to circumvent this adverse property of free allocation. Nevertheless, with targets to achieve net-zero emissions in the foreseeable future, it seems unlikely that there will be sufficient permits available to provide a level playing field for all firms at risk of carbon leakage. A widely discussed measure to address leakage is “border carbon adjustment” (BCA). The intuition behind BCA consists in applying a levy on imports in proportion to their carbon content. Hence, BCA can be thought of as an extension of a domestic carbon pricing scheme to imports (but note that the US is also considering BCA, even though a domestic carbon pricing scheme seems unlikely). However, the logic of applying a carbon price to GHG emissions cannot straightforwardly be applied to other countries. The reason for this is that BCA only affects trade partners' exports. Hence, as shown in Figure 2, it creates incentives for countries to increasingly consume output from sectors facing carbon taxes levied by importers domestically instead of exporting it, or even to shift production to non-export sectors that produce for the domestic market. Therefore, economically optimal design of BCA needs to take into account the magnitude of these substitution effects as well as the carbon intensity of the export and non-export sectors, respectively. BCA is often also discussed as a strategic instrument that could provide a “stick” to punish free-riders for their lack of ambitious climate policies. A recent prominent proposal suggests that major economies should unite into a “carbon club” and jointly adopt BCA against countries without or with laxer climate policies. Even though this idea holds some appeal, it may create the risk of retaliation by countries against which BCA is applied, which might result in a “trade war” and sour relations to an extent that makes cooperation on climate change harder to achieve. In addition, border measures shift some part of the costs of emission reductions to trade partners. If BCA is applied against low-income countries, it risks creating inequitable outcomes by slowing down industrialization and poverty alleviation. Hence, the most preferable option consists in using BCA as a credible warning against free-riders, without actually being forced to carry it out. Finally, carbon leakage becomes less of an issue if mitigation costs are level across regions. Policy makers can help to promote climate measures in other countries by constructively advancing international negotiations. Using the Paris Agreement as an example—under which all signatories commit to submit their “Nationally Determined Contributions” to spell out their envisaged climate change mitigation efforts, in combination with a process to gradually and jointly increase ambition, might be a promising first step in this regard. In addition, access to clean technologies and financial support could provide “carrots” for climate policy laggards to adopt clean energy technologies and production processes. For instance, sectoral agreements to reduce the emission intensity of certain key sectors, such as steel or aluminum, could go a long way toward alleviating leakage concerns. International mechanisms, such as the Green Climate Fund as well as bi- and multi-lateral donors, could be employed to fund such schemes. Another possibility is technical as well as financial support for emerging carbon-pricing schemes, which could help equalize the additional costs accruing to energy-intensive industries across world regions. Legal scholars have extensively discussed the compatibility of border measures to reduce carbon leakage with international trade law, in particular the regulations of the World Trade Organization (WTO). The consensus view in this literature is that BCA could be implemented in accordance with WTO rules, in particular with regard to the requirements of treating foreign producers on equal footing with domestic ones (“national treatment” principle) and not discriminating between trade partners (“most favored nation” principle). Yet, trade law imposes certain constraints on policy design. For instance, it is contested to what extent BCA could be applied in proportion to the actual emissions released in production, as applying different charges on producers from different locations could be regarded as a conflict with the most favored nation principle. Furthermore, trade law would also affect the treatment of domestic producers. Most scholars agree that reimbursing exporters for their costs to reduce emissions in order to ensure their competitiveness in the global market might be problematic because it would most likely be regarded as an export subsidy, which is prohibited under WTO stipulations. The same applies for the free allocation of costless emission permits to domestic producers because it would confer a benefit that is not granted for foreign producers, violating the principle of national treatment. From a political perspective, trade measures are prone to exploitation by vested interests promoting policies that are favorable for their balance sheets. This entails the danger of turning border measures from an environmental to a protectionist instrument. The fact that it is not straightforward to determine which sectors are at risk of leakage opens the door for lobbying by special-interest groups. For the case of free allocation of emission permits in the EU ETS, recent research suggests that a substantial over-allocation to firms that do not face a pronounced leakage risk could be seen as an indication that those firms have indeed been able to successfully influence policy design in their favor. It also seems paramount to apply BCA in ways that avoid tensions with important trade partners. Otherwise, not only trade relations, but also the willingness to cooperate in the international climate negotiations, might be negatively affected. The current discussion of border measures within the EU Green Deal (see Box 1) illustrates the importance of combining the different design features in a way that makes BCA effective and feasible from an economic, legal, and administrative, as well as political, perspective. A pragmatic approach could consist in applying BCA only to a few selected energy-intensive industries that are at risk of carbon leakage based on EU sectoral benchmarks. Such benchmarks are already used as a basis for the free allocation of emission permits and are defined by the average GHG intensity of the best-performing 10% of the installations in the respective sector. Despite some challenges related to the definition of such benchmarks, this approach would be administratively substantially easier than a full life-cycle assessment for all imports, more likely to be in accordance with WTO rules, and politically less contentious, as it would impose lower financial burdens on foreign producers.Box 1Border measures under the EU Green DealAs part of its Green Deal aiming to achieve greenhouse gas neutrality by 2050, the EU has announced the implementation of a “carbon border adjustment mechanism” (CBAM). This CBAM will most likely oblige importers to acquire emission permits from a pool of allowances whose price tracks the one in the EU ETS. The stated goals of this policy include reducing carbon leakage, safeguarding the competitiveness of energy-intensive industries, and generating revenues.The most important design features discussed include the sectors subject to the CBAM, the scope of emissions on which it should applied (i.e., if only emissions directly released in production, or also those related to electricity and intermediate inputs), whether these should be assessed on a life-cycle basis or sectoral benchmarks, and whether there should be exemptions for low-income countries.Because countries with climate policies that are comparable with those in the EU should not be subject to the CBAM, ways to determine equivalence are a further important issue requiring further attention. Finally, regarding the use of revenues, an international climate fund, a domestic mechanism to finance low-carbon projects, or the general EU budget, are options under consideration.Industry representatives have urged the European Commission to gradually phase in the CBAM parallel to a stepwise reduction of freely allocated emission permits (which is currently used to reduce carbon leakage). Such a gradual approach might be especially relevant in view of legal uncertainties. That is, the question whether the implemented CBAM design is consistent with WTO provisions can only be assessed once it is challenged under the WTO Dispute Settlement Body.The European Commission has undertaken extensive consultations with trade partners as well as domestic stakeholders, such as industry associations and environmental NGOs. A decision on the concrete design of the CBAM is expected in June 2021. As part of its Green Deal aiming to achieve greenhouse gas neutrality by 2050, the EU has announced the implementation of a “carbon border adjustment mechanism” (CBAM). This CBAM will most likely oblige importers to acquire emission permits from a pool of allowances whose price tracks the one in the EU ETS. The stated goals of this policy include reducing carbon leakage, safeguarding the competitiveness of energy-intensive industries, and generating revenues. The most important design features discussed include the sectors subject to the CBAM, the scope of emissions on which it should applied (i.e., if only emissions directly released in production, or also those related to electricity and intermediate inputs), whether these should be assessed on a life-cycle basis or sectoral benchmarks, and whether there should be exemptions for low-income countries. Because countries with climate policies that are comparable with those in the EU should not be subject to the CBAM, ways to determine equivalence are a further important issue requiring further attention. Finally, regarding the use of revenues, an international climate fund, a domestic mechanism to finance low-carbon projects, or the general EU budget, are options under consideration. Industry representatives have urged the European Commission to gradually phase in the CBAM parallel to a stepwise reduction of freely allocated emission permits (which is currently used to reduce carbon leakage). Such a gradual approach might be especially relevant in view of legal uncertainties. That is, the question whether the implemented CBAM design is consistent with WTO provisions can only be assessed once it is challenged under the WTO Dispute Settlement Body. The European Commission has undertaken extensive consultations with trade partners as well as domestic stakeholders, such as industry associations and environmental NGOs. A decision on the concrete design of the CBAM is expected in June 2021. Whether carbon leakage will turn into a major concern for climate change mitigation depends on the stringency of climate policies in different world regions. If climate efforts remain at a relatively modest overall level, or if all countries adopt stringent climate targets, leakage seems unlikely to become a major obstacle for climate policy. However, if some front-runners decide to adopt ambitious climate measures while other countries keep lagging behind, there is a real concern that leakage could undermine the environmental effectiveness of, and public support for, these measures. In this case, well-designed anti-leakage policies that are economically, administratively, legally, and politically feasible are clearly required. Hence, the role of BCA as a policy to address carbon leakage is contingent on the question of whether the most important emitters use the post-pandemic economic recovery to “build back better” low-carbon economies, or if some choose to prop up fossil infrastructures that constitute major barriers to future decarbonization efforts. Some of the world's major emitters have announced net-zero targets toward 2050. This opens up the opportunity to form a climate club with common border measures to prevent carbon leakage and provide a motivation for other countries to ramp up their climate ambitions. For instance, if the US, EU, China, and Japan were to apply harmonized BCA, competitiveness concerns for trade flows between these regions would be muted, and other countries would have significant incentives to implement climate measures that are sufficient to grant them access to key market without being subject to BCA. Such a “climate club” can be expected to work best if it does not exclusively rely on sticks to punish non-compliance, but also offers carrots to reward compliance. Access to clean energy technologies as well as financial support could be important building blocks of a strategy to align climate measures adopted on the level of individual nation states with international trade in an integrated global economy. I am grateful to Ann-Kathrin Kühner and Michael Pahle for useful comments and suggestions on an earlier draft of the manuscript.