Impacts of the CBAM on EU trade partners: consequences for developing countries

ABSTRACT This article analyses the impact of the introduction of the Carbon Border Adjustment Mechanism (CBAM) on European Union (EU) trade partners, focusing especially on its potential socio-economic and external consequences for developing and emerging economies (EMDEs). It uses trade data and Multi-Regional Input-Output (MRIO) matrices to investigate the geographically and sectorally uneven distribution of CBAM’s potential impacts. The introduction of the CBAM by the EU is still under discussion. This mechanism, which seeks to reduce the incentives for firms to outsource their carbon emissions and promote a more generalised low-carbon transition, might disproportionally expose some non-EU economies. In absolute terms, Russia, China, Turkey, and Ukraine are the main EU trade partners in CBAM products, and hence the most exposed countries in external and socio-economic dimensions. In relative terms, the degree of exposure of economies that export CBAM products to Europe varies substantially, with many developing economies having more than 2% of their exports and 1% of their production impacted by this measure. East European economies, mainly in the Balkans, as well as Mozambique, Zimbabwe, and Cameroon in Africa, are the most exposed as far as the external dimension is concerned. In socio-economic terms, we can also include Morocco and Tajikistan to the group of most exposed economies. In the end, promoting the substitution of highly polluting technologies with green technologies seems notably easier in Europe than in EMDEs. Many jobs, tax revenues, and export revenues will be lost if the CBAM is implemented without taking into account the specificities of the EU’s trading partners. We discuss options to mitigate adverse consequences on EMDEs. Key policy insights: The CBAM is a logical complementary policy to the EU-ETS, which aims to avoid EU industries outsourcing their production to countries that do not adopt similar levels of carbon pricing. Although most macroeconomic models generally assume that all countries have a relatively high capacity to migrate from one industry to another, evidence shows otherwise. Previous results analysing the impacts of the CBAM might hence have underestimated the consequences for developing and emerging economies. Accounting for rigidities in the production structure, we show that the CBAM may have a regressive impact, therefore requiring careful attention to its institutional design, especially if the objective is to reinforce global climate ambitions in line with the EU’s own decarbonisation strategy. Promoting the substitution of highly polluting technologies with green technologies seems notably easier in Europe than in EMDEs. Many jobs, tax revenues, and export revenues will be lost if the CBAM is implemented without taking into account the specificities of the EU’s trading partners. One possible way to minimise its side effects is to exempt the so-called least developed countries from the CBAM. Rather than an exemption, these countries could also receive targeted support from the EU to reduce their dependence on highly emitting industries, via transfer of technologies, climate subsidies, or concessional lending. The adoption of CBAM-like measures in other developed economies, such as Japan and North America, as well as in wealthier developing countries with the capacity to decarbonise their industries, such as China, would exacerbate the fragilities of emerging and developing economies unless counter-measures are taken.


Introduction
Until now, European industrial sectors exposed to competition from outside Europe were protected through the granting of free GHG emission allowances.However, the European Commission is considering abolishing these allowances because they are considered as suboptimal to achieve decarbonisation in Europe.The competitiveness of exposed sectors vis-à-vis their foreign counterparts that are not subject to similar restrictions could be threatened.This is the main justification for the adoption of the Carbon Border Adjustment Mechanism (CBAM), as empirically discussed by Monjon and Quirion (2011a). 1 Although previous empirical studies have investigated the impacts of similar carbon taxation mechanisms on developing economies, 2 the vast majority of the recent literature on the impacts of the European Union's CBAM focuses on the consequences for EU countries (EU-Commission, 2021;Kuusi et al., 2020).Moreover, most studies pay special attention to the impact on EU firms' competitiveness and the effectiveness of the measure in reducing emissions, rather than the possible macroeconomic imbalances that may emerge in developing and emerging economies.
Recently, a body of literature on the impacts of the low-carbon transition has been discussing the effectiveness of these types of measures in developing countries.The introduction of such measures might generate external and socio-economic constraints for these countries because of their low capacity to adapt their productive structure.Shifting from carbon-intensive industries to green industries (Mealy & Teytelboym, 2020) and technological changes within industries require large investments, which are costly for these economies (Ameli et al., 2021).
In the same vein, this paper addresses the exposure of developing and emerging economies to the introduction of the CBAM.Besides analysing the directly impacted economies and industries using trade data, our study follows Magacho et al. (2023) to evaluate the indirect exposure of employment and wages using Multi-Regional Input-Output (MRIO) matrices.The advantage of these matrices is that they allow us to identify the potential impacts, both within countries and across trade partners, of a demand drop due to trade barriers.Unlike a country's input-output tables, which present domestic sectoral inter-relations and the leakages due to imports, MRIO tables present country-by-product input-output inter-relations.Therefore, besides the potential domestic indirect effects, we can also analyse potential indirect effects due to cross-country trade.
The paper is divided into the following four sections.The next section presents a brief literature review on the economic consequences of the announced CBAM.Even though the announcement is quite recent, many studies have already addressed the effectiveness of this potential measure.The third section describes the methods applied to analyse countries' exposure and the data sources.The fourth section applies the method and presents the results of the analysis.The concluding section discusses possible institutional designs to increase the effectiveness of the CBAM given the findings of the paper.

Main features of the CBAM
The reaffirmation of the EU's climate ambitions under the 'Fit for 55' plan involves a faster reduction in emission quotas.This reduction sends a signal to European producers that they should start making more carbonefficient technology choices.Based on an impact assessment that compares different possible scenarios, the European Commission published a detailed proposal for a CBAM regulation in July 2021 (EU-Commission, 2021).Under the chosen impact assessment scenario, companies importing non-EU goods must surrender carbon allowances equal to the carbon footprint of imported products.In practice, the value of the allowances to be surrendered depends on three parameters: the price of carbon in the EU-ETS, the GHG emission control policies in the country of origin of the goods, and the carbon content of the imported products.
Due to the lack of detailed and reliable data for non-EU countries, a default carbon content of imported goods is computed based on sectoral benchmarks of the 10% least efficient European producers of equivalent goods, although importers have the opportunity to prove a higher carbon-efficiency.Nevertheless, while this facilitates the concrete implementation of the mechanism, it may threaten its capacity to effectively tax imports according to their carbon footprint.Indeed, less efficient producers (those that are far below the 10% worst-performing European producers) will have no incentive to invest in more carbonefficient technologies.
Exporting countries can furthermore claim existing climate policies leading to an equivalent carbon price, and hence reduce the adjustment cost for the importer accordingly.If the originating countries also have a carbon price in the same sectors, then the calculation of the exact adjustment is trivial.It is, however, difficult to assess how this correspondence will be computed when the country of origin has no carbon price but still claims to have other climate policies.Finally, importers are not directly integrated into the intra-European ETS and the carbon price of the EU-ETS is exogenously imposed on them.
The sectors currently covered by the EU CBAM is a shortlist of high-emitting sectors exposed to carbon leakage: four industrial ones (cement, aluminium, iron/steel, and fertilisers) and electricity.Restricting the CBAM to certain sectors and excluding all manufactured or semi-manufactured products considerably simplifies the implementation of the mechanism, as the value chains to be considered within the scope of the CBAM are relatively simple.Furthermore, the EU CBAM only considers direct emissions, and, for many industrial goods, indirect emissions, which are linked to earlier stages in the value chain, account for a significant proportion of the total volume of emissions. 3 The original impact assessment of the European Commission (EU-Commission, 2021) quantifies the effects of each scenario on a number of economic and environmental variables for European economies, mainly using a general equilibrium model called JRC-GEME3.According to the study, the impact of the mechanism on European GDP remains minimal (around a 0.2% decrease by 2030).The effect on production by sector is more marked, as the CBAM has a double impact on production.The strongest impact is on trade flows, with a significant drop in both imports and exports (10% on average in the sectors covered).The European Commission study shows that aggregate impacts on employment are positive, although very low (+0.05% by 2030).The environmental impact of the CBAM, on the other hand, results in a leakage ratio of -29%, which means that a one-tonne reduction in European emissions is accompanied by a 0.29-tonne reduction in non-European emissions.
The impact of the CBAM, however, goes beyond the impact on EU countries.Import adjustments on carbon content applied by industrialised countries are likely to shift some of the burden of emissions pricing to developing or emerging countries (Böhringer et al., 2022).Eicke et al. (2021) propose a first analysis of the CBAM's impact on countries from the 'Global South,' considering two different scenarios (a CBAM focusing only on emission-intensive sectors and one targeting the whole economy).Adding to an important body of country case studies analysing the impacts of the CBAM, especially on China (Bao et al., 2013;Chen & Guo, 2017;Voituriez & Wang, 2011), the authors develop a broad socio-economic and institutional framework to assess the impact of the CBAM.They find that the potentially regressive nature of the CBAM demands careful attention to institutional design, especially if the goal is to enhance global climate ambitions in line with the EU's own decarbonisation strategy. 4 This paper contributes to the literature on the consequences of the CBAM for developing and emerging economies.It also discusses the potential institutional design that would allow the CBAM to foster an equitable and effective low-carbon transition.Its novelty lies in the use of an MRIO table combining detailed sectoral and geographical data.This method allows us to make a systematic assessment of the multidimensional exposure of countries to the implementation of the CBAM, accounting for direct and indirect potential losses.We consider the potential impact on countries' external position, as measured by the capacity of impacted industries to raise foreign exchange, and on socio-economic indicators, such as employment and wages.

Method and data
With the aim of analysing the potential impacts of the CBAM on EU partners, we first compute the share of these products in these countries' exports basket.Secondly, based on the MRIO satellite accounts, we compute countries' carbon emission intensity and compare them with EU countries' emission intensity.Finally, based on Magacho et al. (2023), we estimate the direct and indirect output, wages, and employment related to the production of the CBAM products exported to the EU to assess countries' exposure.

Exports by EU partners
The UN COMTRADE database (UN, 2003) provides data on countries' exports and imports by product and partner.Countries' exports of CBAM products can be computed either based on their exports to EU countries or based on EU countries' imports by partner.Because the EU data is more complete than for many developing countries, we decided to use EU imports to account for countries' exports to the EU.Essentially, a country's total exports to the EU corresponds to the total imports of the 27 EU countries from this country.
There are five products in the current EU CBAM list: Aluminium, Cement, Electricity, Fertilisers, and Iron and Steel.Based on the Harmonised System (HS) Classification, we identify the following products: . Cement: Portland cement, aluminous cement, slag cement, super sulphate cement and similar hydraulic cements, whether or not coloured or in the form of clinkers (2523) .Electricity: Electrical energy (2716) .Fertilisers: Fertilisers, mineral or chemical, nitrogenous (3102) .Iron and steel: Iron and steel (72) .Aluminium: Aluminium and articles thereof (76)

Emission intensity by product
Once the value of exports of these products to the EU (EX EU ) is computed for 2019, it is important to verify whether countries will be impacted or not by this measure.Because the CBAM is not a taxation on products exports, but a fiscal contribution according to products' carbon emissions, countries will not be impacted if their exports to EU countries present very low carbon intensity.
Data on carbon intensity by country and product is obtained from Cabernard and Pfister (2021), who merged the EORA-26 and EXIO MRIO tables into a Regionalized MRIO (RMRIO) table.The advantage of this dataset is that it provides emission-intensity data (measured by CO2eq per USD) for 189 countries and 163 sectors.This provides enough granularity at the sectoral and geographical levels combined.
The emission intensity by product and country is computed as the sum of the emissions of the sectors corresponding to the product under consideration divided by the total output of the product: where e is the carbon intensity, i is each one of the five CBAM products, k is the country under consideration and j is the sector in the RMRIO (which is on the list of CBAM products, as presented in Table A.2 in Appendix).

Measuring countries' exposure to external shocks
We use the EORA-26 MRIO database (Lenzen et al., 2012;Lenzen et al., 2013) for 2015 to estimate the direct and indirect potential impacts on production, wages, and employment.The advantage of using a MRIO model (in comparison to a domestic one) is that besides the indirect impact within a country (due to the supply of inputs for domestic producers), it also accounts for inter-country indirect impacts.If, for example, Country A produces inputs used by Country B to produce aluminium, and this aluminium is exported to an EU country, then Country B is directly impacted, and Country A is impacted indirectly.The total output that is potentially impacted by the introduction of the CBAM (both directly and indirectly) is therefore given by: where ΔX is a column vector of the total change in output (for all sectors and countries), I is an identity matrix, A is the MRIO technical coefficient matrix, and ΔY is the direct impact, which is given by the exports to the EU by country and sector.
Once the total output by sector is estimated, we can compute the total potential impact on wages and employment (both direct and indirect) to analyse the socio-economic exposure resulting from the introduction of the CBAM.To do this, we need to multiply element-by-element the change in total output by the column vector of wages per output (w) and employment by output (n): and where ⊙ denotes the element-wise multiplication, and ΔW and ΔN are the column vectors of the total potential impact on wages and employment, respectively.to Europe concentrated in iron and steel and in aluminium.In particular, Iron and Steel is the most impacted product in Turkey, as well as in the USA, in other BRICS countries (Brazil, India, and South Africa), in South Korea and in Ukraine.
Even though the total impact is relevant, from the point of view of the trade partner, the impact of the CBAM on the economy will depend on its relative relevance (for example, as a share of total exports).As we can see in Figure 1, right panel, the countries that are most impacted in terms of volume are not the same as those that are most impacted in terms of the share of exports, except for Ukraine and Russia.Mozambique is the most impacted economy, as almost 20% of its exports are Aluminium to EU countries.Serbia, Bosnia-Herzegovina, Montenegro, North Macedonia, and Bahrain are also deeply impacted (more than 5% of Serbian and Bosnian exports are CBAM products to the EU, and they represent more than 3% in Montenegro, North Macedonia, and Bahrain).However, unlike in Mozambique the impact in these countries is not always due to a specific product export.In the case of Serbia and Bosnia-Herzegovina, it is due to the export of Iron and Steel, Electricity, and Aluminium, and, in the case of North Macedonia, it is due to the export of Iron and Steel, and Electricity.
Other countries that are strongly impacted relative to total exports are Armenia, Tokelau, Albania, Moldova, and Zimbabwe.In all these economies, exports of CBAM products to the EU represent more than 2.5% of total exports, although the product mix varies significantly from one country to another.In the case of Zimbabwe and Moldova, it is due mainly to exports of Iron and Steel, while in the case of Armenia and Albania, the product mix is more heterogeneous.

Carbon emission intensity by country and product
The analysis of the impact of the CBAM, however, needs to account for emission intensity, as countries' carbon charge will depend on the GHG emissions by product and the country-specific CO 2 intensity.Figure 2 presents the CO 2 equivalent (CO2eq) intensity distribution across EU countries and non-EU countries in logarithmic scale.The dashed line represents the limit of the first decile (10% worst-performing) of EU countries.
For all products but Fertilisers, the carbon intensity tends to be higher in non-EU countries than in EU countries.In the case of Iron and Steel, the carbon intensity of EU and non-EU countries are very similar, with the two curves almost overlapping.In this case, the carbon intensity is relatively low, and most countries are below the average emission intensity of the 10% worst-performing EU countries, which is 1.21 kg per USD.
In the case of Aluminium, the carbon intensity is much lower in EU countries.The worst 10% limit is 0.59 kg per USD, but most of the other countries present an emission intensity much higher than this.In the majority of non-EU countries, the emission intensity is above 1 kg per USD.
Among the CBAM products, Cement and Electricity are the most carbon intensive.Moreover, the difference between EU and non-EU countries is very significant.In most EU countries, the carbon intensity of Cement ranges from 0.5 to 2.0 kg per USD, and Electricity's carbon emissions range from 1.0 to 4.0 kg per USD.The EU worst 10% limit is 4.14 kg per USD for Cement and 6.00 kg per USD for Electricity.

Countries' exposure to the adoption of the EU-CBAM
Because only very few countries outside Europe present low carbon intensity for the products under consideration, most countries can be expected to be impacted by the adoption of the CBAM by the EU.A possible impact is the increase of carbon revenue with no changes in demand or output.This might happen either if the exporter countries absorb all the increase of carbon revenues in their price (and there are no price changes), or if the demand is perfectly inelastic in relation to prices (and hence demand is not impacted by price changes).Appendix A.3 presents the potential fiscal revenues based on countries' exports and emission intensity.
Nevertheless, if prices do change and demand evolves accordingly, the impact of the CBAM will affect not only the sectors directly involved, but also all related sectors through indirect demand effects.The drop in demand will impact the sectors that produce these goods as well as the sectors that supply inputs for the production of these goods.Employers from these industries and from all industries in the supply chain will be affected.
Figure 3 presents the industries potentially impacted, both directly and indirectly, by country.The most impacted countries tend to be the same as before: Zimbabwe, Albania, Armenia, Montenegro, Ukraine, Bahrain, North Macedonia, Bosnia-Herzegovina, Serbia, Moldova, and Mozambique.Nevertheless, one can see that some countries, such as Bosnia-Herzegovina and Serbia, are relatively more exposed due to the indirect effects.This measure accounts not only for the direct suppliers of inputs for the production of CBAM products, but for all indirect suppliers within the country.Therefore, the countries that depend more on productive linkages with the impacted industries tend to be the most impacted.Some countries that were not on the previous list of potentially most impacted economies are now present, because they provide inputs for other countries to export to Europe.This is, for example, the case of Kazakhstan, where the direct impact is low, but the indirect one is very high.
Besides exposure in terms of output, one can also analyse the potential socio-economic exposure of countries to the implementation of the CBAM based on the direct and indirect importance of these industries in generating employment and paying wages.Following Magacho et al. (2023), we estimate the share of employment and the wage bill that might be impacted by the CBAM in EU trade.In Figure 4, the vertical axis shows the share of wages and the horizontal axis the share of employment.The colour of the point indicates the level of Social Protection Coverage, based on data from the International Labour Organization (ILO).As we can see in Figure 4, the share of employment exposed to the CBAM is relatively lower than the share of wages.This indicates that the CBAM exposes jobs that are paid above average.In the case of Mozambique, for example, 6% of wages are exposed, against only 2% of employment, indicating that the 'good jobs' are the ones at risk.
The most exposed countries in socio-economic terms are Moldova (MDA), Mozambique (MOZ), Bosnia-Herzegovina (BIH), Serbia (SRB), North Macedonia (MKD), Ukraine (UKR), Montenegro (MNE), Bahrain (BHR), and Albania (ALB).In these countries, the potential reduction in production puts at risk more than 0.5% of the wage bill and employment.Moldova faces the same problem as Mozambique, with about 2% of employment exposed against more than 5% of wages, indicating that relatively well-paid jobs are the ones that would potentially be impacted.The same problem is faced by Ukraine, even if at a lower scale, where almost 1% of jobs are directly or indirectly related to CBAM products exported to the EU, and these jobs represent almost 2% of the wage bill.
Other countries, such as São Tome and Principe (STP), Armenia (ARM), Russia (RUQ), Georgia (GEO), Turkey (TUR), and Zimbabwe (ZWE), also present an important degree of socio-economic exposure, since more than 0.5% of their wage bill will be impacted.However, in these economies (particularly Zimbabwe), the share of employment at risk is not as high as the share of wages, indicating that the jobs that may be impacted by the introduction of the CBAM in the European Union are few but well-paid.High socio-economic exposure does not necessarily imply high vulnerability.Countries like Ukraine, Bahrain, and Armenia, despite having a significant share of their well-paid jobs exposed to the adoption of the CBAM, are less vulnerable than the other countries discussed above because they have a relatively better Social Protection Coverage (more than 50% of the labour force is covered).Therefore, the potential job losses in the exposed industries in these countries tend to imply relatively lower socio-economic vulnerability.The case of Mozambique, on the other hand, is the most complicated, since about 2% of the jobs (more than 250,000 jobs) and 6% of the wage bill are exposed to the impact, and the country has a weak system of social protection (less than 25% of the work force is covered).

Expansion of the CBAM
One of the main possible political impacts of the CBAM is the incentive it may generate for other countries to adopt similar measures.Because the amount charged as part of the CBAM deduces the current carbon price applied in the country of origin, the introduction of the CBAM may lead EU partners to adopt a carbon price mechanism.We therefore compute in Figure 5 the share of exports that would be exposed in partner countries in order to understand the impact of an expansion of the CBAM to other developed countries.Essentially, we estimate the exports that would be impacted if the following countries adopted a CBAM-type measure: . North American Free Trade Association (NAFTA), which includes the USA, Canada, and Mexico; . Japan and the first group of 'New Industrialized Economies' (JAP + N1), which includes South Korea, Hong Kong, and Singapore; .United Kingdom, Switzerland, and other European Economic Area countries (EEA+), which include Iceland, Liechtenstein, and Norway; .and China (despite not being a developed country).
In terms of total volume of exports, the other BRICS countries (Russia, Brazil, India, and South Africa), as well as Indonesia, Malaysia, Mexico, Australia, and Vietnam would be the most impacted economies.For all these countries, exports of CBAM products to other developed countries and China represent more than USD 2.5 billion per year.The inclusion of NAFTA is especially important for Brazil, which is to be expected since they are on the same continent.The inclusion of Japan and the N1 group is especially relevant for India, Australia, and Malaysia.Finally, the inclusion of China has an amplified impact on Indonesia and Vietnam.In the case of South Africa, all groups of countries but EEA + have relevant impacts (See Figure 5).
In terms of total exports, the most impacted economies are North Korea, Bahrain, Trinidad and Tobago, Zimbabwe, South Africa, Kazakhstan, and Brazil.For all these countries, more than 2.5% of exports would be directly impacted by the adoption of CBAM-type measures.In the case of North Korea, Kazakhstan, and Zimbabwe, the adoption of this type of measure by China is very relevant, whilst in the case of Bahrain and Trinidad and Tobago, it is the adoption by NAFTA that may lead to the highest exposure.

Comparison with existing studies
These results are relatively aligned with Eicke et al. (2021) in terms of key impacted geographies.This study proposed the first analysis of the CBAM's impact on countries from the 'Global South,', considering two different scenarios, mainly a CBAM focusing only on emission-intensive sectors and a CBAM targeting the whole economy.Eicke et al. (2021) developed a broad socio-economic and institutional framework to assess CBAM's impacts, although, unlike us, they did not consider the socio-economic impacts or the indirect drivers of impacts.Their variables of interest are the export structure of countries, their emission intensity, emission reduction targets, and institutional capacities to monitor and report product-based emissions.Additionally, the paper assesses different drivers of exposure and the ability to adapt to a CBAM for the supposedly most impacted countries, namely Mozambique, Bosnia-Herzegovina, and Morocco, which are identified as relatively high-risk.Their findings highlight the importance of the capacity to monitor, report, and verify the carbon content of exporting countries, beyond the mere trade effect.Using an MRIO approach close to ours (but using the WIOD database), Zhong and Pei (2022) focus on Chinese provinces and find that income losses in Chinese landlocked provinces exceed their export losses, contrasting with the pattern for trade-exposed provinces. 5

Policy implications
The policy implications of this analysis are clear.As it stands, this carbon adjustment mechanism, which seeks to reduce the incentives for firms to outsource their carbon emissions and promote a more generalised lowcarbon transition, might disproportionately impact some non-EU economies.Because many of the potentially impacted economies have a low capacity to adapt their productive structure to shift to less-emitting industries or to adopt low-emission cutting-edge technologies, price stimulus such as the CBAM may be less effective in promoting a reduction in emissions in these countries.
These equity concerns could be addressed by returning the revenue from carbon import adjustments to exporters with the aim of reinvesting in decarbonisation, or using it for technology transfer and international climate finance.Addressing equity by exempting some or all developing regions is another possibility, as suggested by (Brandi, 2021), even though this latter solution could, in specific circumstances, undermine the legitimacy of the CBAM by creating new forms of carbon havens (Branger & Quirion, 2014a).These measures, however, are in accordance with the principle of common but differentiated responsibility, which means that developed countries should invest in and take the main responsibility for the development and diffusion of the technologies needed to transform energy-intensive industries (Åhman et al., 2017).In other words, building a development-friendly CBAM, as discussed by UNCTAD (2021), is key to the success of this emblematic part of the Fit for 55 package.
Finally, Russia's war in Ukraine could impact the stability of the CBAM as a new resource for the EU budget for at least four reasons.First, Russia is one of the EU's major trading partners in carbon-intensive goods.GHG emissions embodied in EU final demand from Russia represent around 14% of all GHG emissions imported by the EU.Second, Ukraine ranks fifth in terms of the share of EU imports of CBAM-covered products (5.5%).A postwar Ukraine in need of reconstruction may not trade with the EU at the same levels as before.Third, increased tensions with China because of its support to Russia in the conflict may lead to a similar disruption of its contribution to the CBAM system.Fourth, a persistent increase in gas prices may lead to an increase in the EU's carbon price.While increased carbon prices could mean increased CBAM revenue over time, this is not guaranteed.Any loss of revenue from Russian (and Ukrainian or Chinese) imports would place a heavier burden on exporters from other countries if the EU wanted to raise the same level of revenue.This could also change the way those countries trade with the EU, and how much revenue the EU can expect to raise.
Furthermore, geopolitical tensions could encourage non-EU countries to play the game of trade retaliation.Lim et al. (2021) bring new empirical evidence of the potential retaliation effects from other economies, developing a large set of scenarios generally leading to outcomes detrimental to the CBAM.From a legal perspective, even if CBAM import charges may be WTO-compatible (Tamiotti, 2011), they may not be sufficient to bring many pollution-intensive countries into a carbon-pricing coalition (Böhringer et al., 2022).China and other large non-OECD countries, such as India, Indonesia, and Thailand, currently share the view that the CBAM is a protectionist and discriminatory policy measure.

Limitations
With respect to limitations of this analysis, the methodologies outlined and used in the previous sections are based on adaptations of the Leontief (1936) model, which is commonly used in economics to identify intersectoral dependencies in the event of a demand shock.The model has limitations, including assumptions of constant return to scale, perfect elasticity of inputs or demand, and perfect substitutability among input factors (Dietzenbacher, 1997;Galbusera & Giannopoulos, 2018;Oosterhaven, 1988).Despite these limitations, it can be useful for comparing and contrasting economic structures, determining specific sectors' relative economic/environmental importance (e.g.Antràs et al., 2012;Aldasoro & Angeloni, 2015;Cahen-Fourot et al., 2021;Piñero et al., 2019;Zhang, 2010;Zhang et al., 2020), and assessing the indirect effects of climate policies (e.g.Bastidas & McIsaac, 2019;Chen et al., 2020;Hebbink et al., 2018;Perrier & Quirion, 2018).It is important to note that the model is static in nature and the results should therefore be considered as reflecting current exposure.
The set of products considered here as part of the CBAM is the one outlined in July 2022.Since then, the list has evolved to include hydrogen and 'a limited number of downstream products.' 6Given the undetermined nature of these products and the lack of a hydrogen sector in EORA, we decided to use the July 2022 list.Our results should hence be considered as conservative estimates in this regard.

Conclusion
We analysed the impact of the introduction of the Carbon Border Adjustment Mechanism (CBAM) on the European Union's (EU) trade partners.Using trade data and Multi-Regional Input-output (MRIO) matrices to investigate the geographically and sectorally uneven distribution of the CBAM's impacts, we identify the potential socio-economic and external consequences for developing and emerging economies.
In contrast with studies that use general equilibrium models, and hence rely on assumptions about price elasticities of substitution among countries and technological substitutions, we consider countries' exposure to the introduction of the CBAM based on their exports to the EU and its potential indirect impacts.We identify the current exporters of CBAM products to the EU, and, considering that all exports might be impacted as well as all suppliers of inputs for the industries that produce these export goods, we investigate countries' exposure.These analyses illustrate the degree of exposure of each country, and the potential losses in terms of output, employment, and wages, considering not only the direct impact, but also the impact on all productive chains (indirect impact).The use of MRIO tables allows us to verify that the indirect impact (both within and across countries) plays an important role in determining the degree of exposure of countries.
The degree of exposure of economies that export CBAM products to Europe varies substantially, with many developing economies having more than 2% of their exports and 1% of their production potentially impacted by this measure.East European economies, mainly in the Balkans, as well as Mozambique, Zimbabwe, and Cameroon in Africa, are those where exports are the most exposed.In socio-economic terms, the group of most exposed economies also includes Morocco and Tajikistan.As noted, these results are relatively aligned with Eicke et al. (2021) in terms of key impacted geographies.
The potentially regressive nature of the CBAM requires careful attention to institutional design, especially if the goal is to enhance global climate ambitions in line with the EU's own decarbonisation strategy.Brandi (2021) argues that least developed countries (LDCs) should be exempted from the CBAM.Instead, they should receive targeted support from the EU.The effectiveness of the CBAM relies on a broader set of development policies to support the most exposed countries in the formulation of their own carbon neutrality strategy.Finally given the uncertainty surrounding the war in Ukraine, future trade relations with Russia, Ukraine, and China, and the overall CBAM revenue structure, it is difficult to compute the exact amount of potential CBAM revenue which would be impacted, but it is safe to say that trade disruptions between Russia and the EU would negatively impact the EU's budget.

Notes
between products from different countries, are a key parameter in explaining the different results obtained by the various carbon leakage models. 7he impact study uses a carbon price of 47 euros per tonne in 2030.It would be legitimate to question this level and its compatibility with meeting climate change objectives.The Carbon Pricing Leadership Corporation reminds us that it is necessary to have reached a carbon price of at least 50-100 dollars by that time to meet the commitments of the Paris Agreement.The price per tonne of CO2 under the CBAM will be aligned with the price of a tonne of carbon in the ETS, which fluctuates with supply and demand, the current price in the ETS system being close to 60 euros per tonne.
Finally, the risk of circumvention through resource shuffling could seriously undermine the effectiveness of the scheme.Resource shuffling consists in artificially reallocating emissions between different production factors, sectors or companies.By manipulating carbon accounting in this way, a state or a company could claim to have reduced the carbon content of goods exported to the EU without having made any real changes to the production process.These manoeuvres are both difficult to anticipate ex ante and difficult to measure ex post.(Stede et al., 2021) estimate a risk of resource shuffling of 50% to 80% for the aluminium sector.
In the end, as noted by Böhringer et al. (2022), the CBAM raises more concerns about legal feasibility (see also Tamiotti, 2011) and practical implementation than about fundamental theory.If it is perceived abroad as a form of protectionism, it could lead to trade conflicts.At the same time, it has also been proved that internationally traded goods are biased towards high carbon products because of differentiated import tariff or non-tariff barriers.As already shown by Monjon and Quirion (2011b) and recently confirmed by Mörsdorf (2022), although a CBAM would be effective at limiting leakage, a decrease in European production of GHG-intensive products is still to be expected.Industries that consume cement, aluminium and steel would pay more for these goods under a CBAM.

A.1.2. Impact of the CBAM outside the EU
The impact of the CBAM goes beyond the impact on EU countries.One of the aims of the CBAM is to foster global transition dynamics.It is thus crucial to analyse its potential effects on the rest of the world.The CBAM will likely change the price dynamics for some countries and disproportionately impact some economies.As stated in (Droege, 2011), the UN emissions inventory system relates to the point of production, not consumption, following the 'polluter pays' principle.Policy makers need not only to determine the environmental integrity, the administration and the political credibility of border measures, but also to consider that any such measure, even if it aims at supporting global emission reductions, could disrupt the international climate negotiation processes.
Indeed, import adjustments could have a strong burden-shifting effect, as suggested by (Böhringer et al., 2022).Import adjustments on carbon applied by industrialised countries such as the EU countries are likely to shift some of the burden of emission pricing to developing or emerging countries.Exporting countries subject to the CBAM could suffer a loss in export revenues, employment and wages, or in fiscal revenues, as we confirm in this paper.As explained by (Böhringer et al., 2022), this outcome may potentially be at odds with the principle of common but differentiated responsibilities articulated in the United Nations Framework Convention on Climate Change, and reaffirmed in Article 4(3) of the Paris Agreement.As a counter to that argument, the CBAM would also contribute to reallocating the responsibility of emissions to consumer countries (in this case the EU) and mitigate the 'ecologically unequal exchange' of globalisation.
After a model-based comparison of the impact of the CBAM and an extension of the EU ETS to the power sector, Pató et al. (2022) conclude that expanding the geographical scope of the EU ETS is a more effective climate policy tool than the CBAM.The authors argue that it would reduce emissions, while the CBAM would not.They also find that expanding the ETS yields more revenue to exporting neighbouring countries facing higher-than-average challenges to change their fossil-heavy power systems.Furthermore, while Evans et al. (2021) consider that replacing free allocation by an import-only CBAM would weaken the competitiveness of EU producers in foreign markets, Pató et al. (2022) consider that the CBAM is a competition policy tool that provides a level playing field for producers inside and outside the EU.Eicke et al. (2021) propose a first analysis of the CBAM's impact on countries from the 'Global South', considering two different scenarios, mainly a CBAM covering only emission-intensive sectors and a CBAM targeting the whole economy.Adding to an important body of country case studies analysing the impacts of the CBAM, especially on China (Bao et al., 2013;Chen & Guo, 2017;Voituriez & Wang, 2011), the authors develop a broad socio-economic and institutional framework to assess the impact of the CBAM.Their variables of interest are the export structure of countries, their emission intensity, emission reduction targets, and institutional capacities to monitor and report product-based emissions.Additionally, the paper assesses different drivers of exposure and of the ability to adapt to a CBAM for Mozambique, Bosnia-Herzegovina, and Morocco, identified as relatively high-risk.It highlights the importance of the capacity to monitor, report and verify the carbon content of exporting countries, beyond the mere trade effect.Using an MRIO approach close to ours (but using the WIOD database), Zhong and Pei (2022) focus on China and find that income losses in Chinese landlocked provinces exceed their export losses, contrasting with the pattern for trade exposed provinces.

Figure 2 .
Figure 2. Distribution of emission intensity, by product (2015).Source: Authors' elaboration based on COMTRADE and RMRIO data Scales are in logarithm; dashed lines represent the CO2eq emission intensity of the 10% worst-performing EU countries.

Figure 4 .
Figure 4. Exposure to the CBAM in terms of wages and employment, by country (2015).Source: Authors' elaboration based on COMTRADE and EORA data.Scales are in logarithm.