The definitive phase of EU CBAM begins in January 2026, but benchmarks for steel products have still not been formally agreed. An issue is that the facility-level benchmarks used under EU ETS are not directly applicable for CBAM, which needs product-, or customs code-, specific benchmarks.
Therefore, we used CRU’s Steel Cost Model to determine the product-level free allocation resulting from the facility level benchmarks under EU ETS for individual integrated mills in the EU and further used these to identify a ‘fair’ benchmark for imported steel under CBAM.
We suggest a CBAM benchmark set by the median free allocation for EU mills would ensure imports receive the same average level of free allowances as EU mills do under EU ETS. However, recently ‘leaked’, but unverified, benchmarks for steel seem to be closer to our 90th percentile measure for EU mills. This implies a distinct advantage for imports.
Having said that, it seems the ‘leaked’ benchmarks are slated for use specifically when assessing the CBAM charge using ‘default’ emission values, which we believe are elevated relative to actual emissions. High default values may more than offset any advantage from a high benchmark in many cases.
No benchmarks have yet been officially released (or ‘leaked’) for use when assessing the CBAM charge using actual, verified emissions (n.b. values given in the leaked documents are very low and unrealistic, which means they are likely to be ‘placeholder values’ in the draft document).
We have determined fair benchmarks for the full range of downstream flat products (e.g. hot rolled, pickled and oiled, cold rolled, EG, HDG and OCS) as well as slab and reversing mill plate and, coupled with our quarterly carbon price forecast for next year, these may be of immediate use to EU buyers that are negotiating import deals today for delivery in the new year.
If you want to talk through how we have determined steel benchmarks or to understand our carbon price forecasting, get in touch with us,, we’ll be happy to go through our thinking.
