Default values in the UK’s CBAM could hinder its climate ambitions

The application of national average emission intensities

Like the EU’s, the UK’s default values are likely to rely heavily on estimations of national emission intensities. In both cases, given that emission intensity can vary significantly across firms, deriving defaults from a country’s average emissions for each CBAM-covered good, potentially favours less-efficient producers within a country at the expense of more efficient ones. To illustrate this, consider Figure 1 which plots the distribution of energy intensity across firms, where we take energy intensity as a proxy for emissions intensity (for which there is no data). Figure 1 plots the energy intensities of Indian firms in the basic metals industry in 2013, based on data from the World Bank Enterprise Survey (WBES). 4

This figure is likely to be fairly representative of the distributions of emission intensity across firms relevant to the UK CBAM. First, we expect that firms’ energy intensities follow a similar pattern to their emission intensities within a specific industry, assuming similar production technologies are used across firms. Second, the basic metals industry includes the production of steel and aluminium, which are significant components of the UK CBAM, accounting for 80% of the UK’s total CBAM-good imports in 2023. Last, India played a key role in the UK’s CBAM imports, exporting £610 million worth of CBAM-goods (3% of the UK’s total CBAM-good imports) to the UK in 2023.5

In Figure 1, energy intensity is calculated as the ratio between the total annual expenditure on fuel and electricity and the value of total annual sales, following the methodology of Montalbano et al. (2022).6,7 The distribution shows a long tail to the right, with the median value around 0.025 and the mean value at 0.04 (indicated by the dashed lines). These highlight a significant disparity in energy intensity among basic metal-producing firms in India. Once a default value was set, there would be a truncation effect - the firms with an emission intensity above the default values may choose to rely on them instead of providing actual data. For example, if the mean is set as the default in Figure 1, 125 out of 343 firms may use default values. Thus, emissions of 22.5% of total industrial sales might be underpaid. Consequently, the reliance on default values by less efficient producers could allow them to avoid facing a price comparable to that encountered by local UK producers.

Further, applying country-specific estimates also does not account for the emission differential between exporters and non-exporters. Richter and Schiersch (2017) find a negative relationship between export intensity and emission intensity, suggesting exporting firms are less CO2-emission intensive than non-exporting competitors.8 If the information on export production is available, the averages can be calculated over only exporting firms which would better represent exporter’s emission intensity. However, in terms of exports, emission intensities can also vary across destination countries. Liu et al. (2024) report a differential CO2-emission intensity between Chinese exports destined for developed countries and those destined for developing countries.9 This might suggest “emission shuffling” whereby producers send their cleaner production to markets with stricter environmental regulations while directing dirtier production to markets with less stringent regulations. The incentives for shuffling would become greater with the introduction of the EU or UK CBAM. If we could accurately identify the emissions embedded in a firm’s exports to the UK/EU, that would seem more in line with the spirit of the CBAM, but separating out export production destined for different countries and the related emissions at the firm level is very challenging.10 Moreover, the default value is for circumstances where we do not have verified export-firm data, so isolating exports to particular destinations seems likely to be even more complex.

The application of common default values across countries

Although the UK Government “is minded” to apply a similar method of calculating default values as the EU, the consultation document clearly shows a departure from the EU in terms of what the default is applied to. Whereas the EU intends to apply country-specific default values where it has data, the UK “is minded” to apply the same default to all countries: “The government considered designing jurisdiction-specific default values. However, this was deemed infeasible by 2027.” This indicates that the “global average” will be applied to all importers, regardless of their specific country conditions like the technology level, even when more specific data is available.

This method, described in the consultation document, seems perverse. The method indicates that the “global average” requires weighted national emission intensities to be calculated. However, if these national emission intensities were known, it raises the question of why the UK would apply the same values to all importers instead of adopting the EU’s proposed approach, which tailors default values to specific countries?

As might be anticipated, there is a significant variation in emission intensities across source countries, even at a very disaggregated level, as indicated in the above JRC report. For instance, Ferro-nickel, which is one of the goods covered by both the EU CBAM and the UK CBAM, classified under the CN code 72026000, exhibits varying emission intensities across countries. These intensities range from 4.49 tonnes of carbon dioxide equivalent per tonne (tCO2e/t) for Brazil to 9.76 tCO2e/t for Serbia, with a weighted average of 6.26 tCO2e/t. Notably, 3 out of 9 countries/regions report an emission intensity above the average one.11,12

To illustrate the negative effect of applying the same default values to all countries, we utilised the default values estimated in the JRC’s report for four UK CBAM-sector imports (iron & steel, cement, fertilisers and aluminium) in 2023.13 Assuming that imports from countries with higher emission intensities than the average values would resort to default values instead of providing actual data, we computed the underestimated emissions (the disparity between the national emission intensities and the average values) from these countries. This analysis was conducted under the scenario where the UK uses the EU’s estimations of the default values but applies the average values to all countries within each category.

The JRC’s samples of key trading partners and producers is based on EU imports, but it also offers very substantial coverage of UK imports of CBAM-goods: 86%, 93%, 82% and 88% of the UK's CBAM-imports in iron & steel, cement, fertilisers, and aluminium in 2023, respectively.

Our analysis reveals that the underestimation of emissions could exceed 1 million tCO2e in the iron & steel sector and 1.5 million tCO2e for the combined four sectors, as outlined in Table 1. Notably, related imports constitute more than 10% of the UK’s total CBAM-imports. Most of them are concentrated in cement (91% of its imports) and aluminium (11% of its imports) sectors. Therefore, this underscores the potential adverse impact on market fairness for both domestic producers and producers from other countries with lower emission intensities, especially within certain sectors.

Furthermore, the underestimation weakens the effectiveness of the UK CBAM and impedes the UK’s efforts to achieve net zero, although the numbers are minor compared to total UK GHG emissions - over 502 million tCO2e in 2021.14

Table 1: UK’s underestimated emission from CBAM-imports in 2023

Notes: the numbers come from the author’s calculations.

‘Production vs. trade’ weights

Similar to the EU, the UK proposes to calculate the global average values using each country’s production volume as weights. The key to calculating the averages is to find values that accurately reflect the actual emission intensity of imported goods. The appropriateness of the methodology also depends on the source countries to which the averages will be applied.

This approach, using production weights, appears less reasonable, given that the defaults apply even to imports to the UK for which national data is available. For instance, consider two countries producing a CBAM good. Country A produces a large quantity with minimal exports into the UK, while country B produces less but exports more into the UK. By using production weights, country A would significantly influence the average. However, this average would not accurately reflect the average emission intensity embedded in the UK’s imports. In contrast, using import weights would yield a much fairer average among all importers.

However, if the UK intended to apply the defaults only to countries that are not its key trade partners or key producers, taking a similar approach to the EU’s, i.e. using production weights, may be more appropriate. The challenge is to estimate the emissions of a small import source to the UK/EU on which you have no data, and here, assuming it has ‘typical’ global emission intensity may seem more reasonable. The global averages would be applied to countries whose exports of CBAM goods to the UK or the EU are very limited, probably below 10% of total imports, according to EU data. Therefore, using import weights would not provide a fair average for them. Conversely, averages weighted by production would more accurately represent the global average level. This approach could potentially offer a better indication of emissions embedded in these imports to the UK.

Note that the UK CBAM consultation document exclusively references the key trade partners of the UK. Additionally, we include ‘key producers’, as mentioned by the EU, as averages weighted by production will represent the global average level only if key producers have been also accounted for in the computation.

The percentage-based mark-up method

The UK Government proposes a percentage-based markup as a further option for default values. Specifically, it is based on the above “global average” and involves “a percentage-based markup” as outlined in the consultation document. Similarly, the EU has also announced that a new set of default values to be implemented from 2026. These values will be increased by a proportionately designed markup based on the average emission intensity in default values for CBAM goods in the transitional EU CBAM period. However, there is no further information regarding the design of markups.

An appropriate multiplier applied to the estimated intensities could offer a viable solution to sufficiently reduce reliance on default values without causing significant trade diversion by making default values prohibitive. However, the most challenging aspect lies in determining the appropriate markup. Setting appropriate markups relies significantly on understanding the difference between the default values and the actual emissions from those firms using default values.

The minimum threshold setting

In the consultation document, the government proposed that “a person will not need to register or account for CBAM if the total value of their CBAM goods falls below a minimum registration threshold of £10,000 over a rolling 12-month period.” and that “the threshold will exclude around 60 per cent of potential registrations, while retaining more than 95% of emissions.” The trade-off implied by this threshold cannot be assessed accurately without a little more detail – specifically, would be the shares of emissions retained for other potential thresholds (say, £7,500, £12,500 and £15,000).

The minimum threshold for CBAM goods serves as a form of default – if you are a small importer, we’re willing to behave as if the default were zero emissions. An alternative to this threshold, based on emissions, would offer a more direct and fitting approach, aligning with the objectives of the UK CBAM to address carbon leakage risk and promote decarbonisation.

This methodology avoids the significant variability in emissions across CBAM goods. For instance, applying the JRC’s estimates of country-specific emission intensities and the UK’s CBAM imports in 2023, we calculate the emission embedded in each CBAM good valued at £10,000. It can range from 0.001 to 233.89 tonnes of carbon dioxide equivalent per £10,000 (tCO2e/£10,000). To be specific, if the £10,000 were devoted to bars and rods, hot-rolled, in irregularly wound coils, of iron or non-alloy steel (hot-rolled bars of iron and steel product with the Combined Nomenclature code 72139990) from Switzerland, emission of 0.001 tCO2e would be implied, but if it were devoted to aluminium foil, backed, of certain thickness (aluminium foil with the Combined Nomenclature code 76072099) from Egypt, emissions of 233.89 tCO2e would be implied. This range becomes even more pronounced - 0.01 to 520.81 tCo2e/£10,000, when average estimates of emission intensities are applied uniformly across all source countries.15 We excluded the outliers in this exercise, 75 out of 7,673 observations (representing the top and the bottom 0.5%) in terms of kilogram per £10,000 (kg/£10,000).16 Otherwise, the emission range would be even broader.

The application of this methodology would heavily rely on default values because importers who come under the threshold are more likely to be small enterprises, thereby relying on default values rather than submitting actual data. Thus, default values would frequently serve as determinates for CBAM registration. In this context, establishing fair default values is essential in defining the minimum threshold and becomes even more critical.

An information-collection stage

The establishment of default values is highly dependent on the availability of data. While the UK Government has expressed its intention to explore various data sources, including the data compiled by the JRC, the inclusion of two additional sectors - ceramics, and glass, in the UK CBAM compared to the EU’s requires additional new data.

The data submitted by importers is also crucial for establishing default values. They will require time to adopt the reporting process, for example, all the paperwork and the new reporting system. Similarly, the government will need time to analyse the submitted data to better set default values. Therefore, a well-structured information-collection stage would be beneficial.

Given that establishing default values relies heavily on actual data, an information-collection stage is desirable (unless the UK CBAM aligns with the EU’s). Without such a phase, gathering the relevant emission data could pose a significant challenge in the setting of fair default values for the UK CBAM.

Conclusion

Setting the optimum default values for emission intensities in the UK CBAM is a critical step to ensure its effectiveness and fairness. Our analysis has highlighted the complexities involved in determining these default values, considering factors such as the emission-intensity differences across producers, and the production/import weights in computing the averages.

We recommend establishing specific default values for key trade partners and major producers of CBAM goods, and applying global averages weighted by production to other importing countries. This approach can more accurately reflect the emission intensities of imports, thereby maintaining the environmental integrity of the CBAM without being punitive.

Additionally, we propose an alternative approach to the minimum threshold, based on emissions, which is both more direct and more appropriate. A robust information-gathering phase is essential to refine these values and address the technical challenges that are likely to arise. By implementing these strategies, the UK can better manage carbon leakage risks and promote global decarbonization efforts.