In this report, “carbon” is used as a synonym for GHG emissions, in line with widespread use in the area of GHG metrics. However, the term “carbon” is misleading, as the GHG metrics do not cover carbon dioxide (CO₂) alone but also other GHG within the meaning of the Kyoto Protocol, such as methane (CH₄) and nitrous oxide (N₂O).

The various GHGs are measured by their GHG or warming effects in CO₂e or tCO₂e.

Annual global GHG emissions are around 50 billion tCO₂e. 2 In Germany, they amount to around 760 million tCO₂e. This means that average per capita emissions in Germany are just under 9 tCO₂e.

To calculate GHG metrics for Bundesbank investments in banks, GHG data derived from the banks’ disclosures (e.g. sustainability reports) are used. The Bundesbank obtains relevant data from the ESG data provider ISS ESG. When calculating GHG metrics, the Bundesbank does not use data on GHG emissions modelled or estimated by data providers. 3  

In principle, the aim is to ensure that portfolio holdings and GHG data relate to the same year, data availability permitting. However, for the reporting dates of both 31 December 2023 and 31 December 2024, the GHG metrics are based on data on banks’ GHG emissions in 2023. These were the latest available GHG data when this report was prepared, which is why they were also used for the portfolio reference date of 31 December 2024. For next year’s climate-related disclosures, the plan is to retroactively adjust these calculations using GHG data for 2024. Following the same principle, the calculations as at 31 December 2023 reported in this document have been adjusted to reflect the GHG data for 2023 (having been based on the 2022 data in last year’s disclosures).

The GHG data on sovereigns used in this report are mostly obtained from the United Nations Framework Convention on Climate Change (UNFCCC) data interface, where the UNFCCC lists GHG data recorded by sovereigns and reported to the UNFCCC. In cases where no recent GHG data are available for a given sovereign in the UNFCCC data interface, GHG data from the World Resources Institute (WRI) modelled for this report are used. The data are obtained using the production-based, or territorial, measurement approach: sovereigns are assigned those GHG emissions that occur within their borders or in their jurisdiction. 

Production-based/territorial GHG emissions are also used for sub-sovereigns or federation members/regions. The respective national authorities (e.g. statistical offices) serve as data sources.

GHG data based on the consumption-based measurement approach are not yet included in this report. This approach draws on foreign trade data. Unlike in the production-based approach, sovereigns are not allocated GHG emissions from exported goods but are additionally allocated GHG emissions from imported goods. However, a number of consumption-based methods have so far led to – in some cases – large discrepancies in GHG levels. This report therefore focuses on the well-established production-based approach. 

In principle, the aim is to ensure that portfolio holdings and GHG data relate to the same year, data availability permitting. However, GHG data on sovereigns have longer lead times than GHG data on companies. At the time of reporting, the majority of the latest available GHG data refer to 2022 or 2023. In most cases, then, the GHG data merely serve as a provisional basis for the GHG metrics as at the portfolio reference dates of 31 December 2023 and 31 December 2024. To nonetheless provide a meaningful picture of developments in sovereigns’ GHG profiles over time, longer periods are considered for the GHG metrics (WACI, total carbon emissions and carbon footprint) than for amounts that are not invested in sovereigns.

Measurements of sovereign GHG emissions often differ in whether they include the land use, land use change, and forestry (LULUCF) sector. For example, LULUCF can be a major – and sometimes the largest – source of GHG in densely wooded regions given deforestation. By contrast, reforestation may lead to negative emissions, resulting in an improved GHG profile. LULUCF then constitute GHG sinks.

LULUCF thus play an important role in many national and international climate objectives. However, in view of the complicated accounting of LULUCF, climate researchers still have reservations concerning the quality and comparability of such data. 

The GHG metrics in this report are shown both with and without LULUCF.

The methods for calculating the total carbon emissions and carbon footprint are not as well established for investments in sovereigns as they are for investments in companies. Methodological challenges notably include adequately capturing which sovereign shares and thus which GHG emissions are financed by investments in the portfolio. 

Early climate-related disclosures often took sovereign debt as the reference variable. Government bonds (held in a portfolio) do indeed account for a part of sovereign debt. However, sovereigns and their GHG emissions are not financed through sovereign debt alone. Under this approach, the financed emissions would therefore be overstated.

The PCAF recommends using GDP adjusted by PPP as the reference variable instead. Unlike sovereign debt, GDP adjusted by PPP is clearly related to a sovereign’s GHG emissions. However, the relationship between government bonds (held in a portfolio) and GDP adjusted by PPP is unclear.

Both approaches therefore have substantial downsides. Despite these reservations, this report includes the total carbon emissions and carbon footprint based on both approaches in order to contribute to the discussion of methodological developments and to identify differences in results.

Production volumes of coal, crude oil and natural gas are measured by energy content in terajoules (TJ), 4 and are set in relation to the size of the economy (GDP adjusted by PPP), analogously to the WACI methodology. Weighted averages of the intensity of coal, crude oil and natural gas production volumes are calculated for sovereign investments on this basis.