EU CBAM: Well intended is not necessarily well done

The devil is in the details: The EU carbon border adjustment mechanism (CBAM) will not only become more expensive than anticipated but will likely fail to restore a level playing field for industries heavily exposed to carbon pricing, first because it is an invitation for greenwashing. For setting the “tariff” of the CBAM, a company can use an individual assessment of its emissions instead of the default emission intensities provided by the EU. This is in particular problematic with indirect emission (e.g. electricity): Many foreign companies will find ways to completely attribute their use of green electricity – even if only a small share of their overall consumption of electricity – to the goods exported to the EU. This can easily be done if so called “power purchase agreements” (PPAs) are in place. The result is then merely the reallocation of existing renewable electricity to products imported into the EU, using the remaining brown electricity to produce goods for non-EU markets. Within the EU, producers don’t have the option to use this set-up to evade carbon pricing obligations for products they sell outside of the EU, unless they relocate their production for non-EU customers outside of the EU.

In addition, the focus on keeping the free allocation of certificates in place while trying to secure WTO compliance undermines the possibility for reimbursing the carbon levies of exports, ultimately shifting the competitiveness and carbon leakage issues just one step down in the value chain. Ideally, the EU CBAM would not only need to provide a level playing field for products sold on the EU market, but also for EU-based producers selling their products on non-EU markets. This would be possible with a focus on levying a carbon price only on products locally sold in the EU combined with a much broader product base, which includes important complex downstream products beyond the currently covered basic goods.  Such an approach would give domestic producers the right to reclaim the carbon levies when exporting a product. It is well advisable for future revisions and a broadening of the EU CBAM product base to explore the options provided by the possibility of pricing carbon at the product level but determining the related emissions and collecting levies further upstream. This, for example, is the way the “national German emission trading system on fossil combustion for heat and transport” works, collecting the fees at the wholesale level while aiming at downstream emissions.

Much of the concerns regarding the WTO compliance stem from the free allocation of EU emission certificates.  Depending on the sector, more or less of the EU emission trading system (EU ETS) certificates are allocated free of charge according to best-in-class benchmarks. This free allocation is supposed to be phased out during a ten-year period following the start of the regular CBAM in 2026.  It remains to be seen if the WTO deems the suggested reduction of EU CBAM levies in proportion to the share of free allocations a fair practice. The focus on keeping the free allocation of certificates, however, has two drawbacks: It limits the flexibility of reforming the EU ETS towards a level playing field. And it excludes carbon levy reimbursements for exports as the WTO would regard such a combination as an unfair trade practice. The consequence:  the competitiveness and carbon leakage issues is shifted just one step down in the value chain: if higher carbon related costs are passed on, e.g. from steel makers to EU car manufacturers, the car manufacturers in turn might choose to relocate part of their production outside the EU in order to gain access to cheaper steel for their non-EU automobile customers.

In its current form, the EU CBAM also lacks common actions to support the industries included. The European Commission acknowledges the risks of carbon leakage in their assessment. However, the expected loss of employment of just over 1% in CBAM sectors – and only a minimal loss of employment in downstream sectors – might be a little too optimistic. Also, the downstream losses are very unevenly distributed between sectors. Some sectors, such as construction, even gain in the EC assessment. The main losers are other non-ferrous metals with -0.9% output loss, the transport equipment sector with -0.4% (including the car manufacturers mentioned before), crops with -0.3% and chemicals as well as other equipment and consumption goods with around -0.2% in 2030.   

A certain way to limit the negative impact could be to ensure that the European “green” versions of basic goods covered by the EU CBAM are cost-competitive “at the source” (i.e. without free allocation of certificates or reimbursements of carbon levies) compared to their foreign brown counter-parts in non-EU markets. Member states such as Germany have announced that they will establish mechanisms that aim at achieving exactly that. The policy instruments of choice for, say, green steel, aren’t limited to investment subsidies for blast furnaces that operate with hydrogen. They could also include so-called “carbon contracts for difference” (CCFD) that subsidize the green steel when it is sold, thus reducing not only the CAPEX of the underlying green investments but also subsidizing their OPEX to a competitive level. While the EU vaguely promotes the use of CCFDs in this respect, a coordinated and adequate EU strategy is still missing.

Overall, the EU CBAM could lead to a higher-than-expected burden for importers, especially Russia, Turkey and Ukraine. Previous analyses of the EU CBAM burden were based on the assumption that the product emission intensities are oriented towards average EU emission intensities while in the current version of the EU CBAM, default emission intensities are calculated from actual average emissions intensities of a respective country’s products plus an ominous markup that will be determined in a yet to be specified procedure.  Or, alternatively, default emissions are calculated from the worst 10% of similar EU producers. Default values for electricity are calculated with a different approach, but we will come to that later.

Figure 1 – Exemplary assessment of EU ETS benchmark installations in the carbon steel production

As displayed for the carbon steel benchmark installations (EAF: electric arc furnace) in Figure 1, the emissions vary considerably over different production sites in the EU. The analysis shown is originally used in the benchmarking process to determine the free allocation of emission certificates from the best-in-class emitters. Simply guesstimating from the graph indicates that the 10% of installations with the highest emissions intensities have an emission intensity of around 0.7 compared to the average emission intensity of around 0.3.

Table 1 – Exemplary assessment of EU ETS benchmark installations

Similar benchmark installation curves are available for 52 products,  and Table 1 summarizes our analysis of the relevant EU ETS benchmark installations.  Clearly, these are only indicative as the EU CBAM will provide its own assessment of benchmark installations in coherence with the list of goods covered. Nevertheless, the available data suggest that a focus on the “worst in class” increases the carbon price around 70% on average, with steel and fertilizers facing increases of more than 90% compared to what was previously anticipated.

Figure 2 – Countries most severely affected by EU CBAM for iron, steel (upper graph) and cement (lower graph)

Figures 2 and 3 give updated estimates for the range of the burdens to be expected for the most affected countries. The calculations from the 2021 SWP study “A CO2 border adjustment for the EU Green Deal”  by Susanne Dröge serve as a benchmark for the previously expected burden based on average EU emission intensities while the darker shaded column of a column pair indicates the adjustments according to the new 10% worst emitters benchmark in EU CBAM. The “CN product classifications” in the EU CBAM are not identical with the “NACE sector classifications” used here so the absolute values on the left axis deserve a further assessment. That neither changes the validity of the relative differences between the previous and the new assessment, nor of the surcharge on the right axis. The calculations in Figure 3 differ in so far as default intensities for electricity are supposed to be based on the respective country’s “average CO2 emission factor in tons of CO2 per MWh of price setting sources”. This particularly benefits Albania, which has an actual carbon intensity of electricity close to zero (thanks to its vast use of hydropower), thus much lower than the average EU carbon intensity employed previously by SWP.

Figure 3 – Countries most severely affected by EU CBAM for electricity

These calculations, however, come with a caveat: As producers can use the certified actual emission intensities of their production process instead of the default values, the actual EU CBAM payments will be lower than the ones calculated here with the EU CBAM default values. Setting the default benchmark by the 10% worst benchmark installations represents a strong incentive to use the option of certifying one’s own production process. It is to be seen whether producers who export into the EU manage to reach or even undercut the average EU emission intensity. If they do so by greenwashing, the EU could lose another important opportunity along its decarbonization path.

Economists tend to be huge fans of a carbon border adjustment mechanism as it equalizes emissions reductions incentives beyond national borders and eliminates the distortions to the global level-playing field caused by regional carbon pricing, at least in theory. Unfortunately, we don’t live in a theoretical world and the devil is in the details (see Appendix). A German proverb states that “well-intended” is the opposite of “well-done’”. The coming years will show how much of the good intentions can be transferred into a well-done regulation. As with the Emissions Trading System, the EU is exploring uncharted territory with the CBAM, which should at least have a better start and hopefully become a similar success, even if it has scope for improvement.