Transparency data

UK Export Finance Annual Report and Accounts 2023 to 24: Task Force on Climate Related Financial Disclosures (TCFD)

Updated 31 July 2024

UKEF Financed Emissions: Basis of Reporting  

1. Context  

Since 2022 UKEF has annually published its portfolio-wide financed emissions estimates. As the first Export Credit Agency (ECA) to publish its financed emissions, UKEF has developed a bespoke double disclosure methodology, in the absence of a Partnership for Carbon Accounting Financials (PCAF) methodology for ECAs. UKEF’s bespoke methodology is based on the dual “follow the money” and “follow the risk” principles articulated in the PCAF Standard, reflecting the unique role of ECAs[i].

UKEF produces these estimates on a best-efforts basis, however the considerable outstanding methodological and data quality issues faced by UKEF and prevalent across the financial industry mean that these metrics are still evolving and not yet at the same standard as more traditional financial metrics. Estimates are subject to a number of limitations, including limited verification of data accuracy, which are described in more detail below. Publishing estimates in this environment carries the risk that metrics will need to be updated to reflect subsequent advancements in climate-related data, modelling, and methodologies. We publish PCAF-aligned data quality scores alongside our estimates, to give an indication of the quality of data sources used.

We continue to work with our peers to support the development of a global financed emissions disclosure standard applicable to common ECA product types such as guarantees and export credit insurance and to address the data quality issues described below.

2. Methodological approach

UKEF’s approach to producing financed emissions estimates is informed by The Global GHG Accounting and Reporting Standard for the Financial Industry[1], developed by PCAF. PCAF is a global partnership of financial institutions working to develop and implement a harmonised approach to assess and disclose the greenhouse gas (GHG) emissions associated with their activities. In 2023 UKEF became a PCAF signatory.

PCAF provides accounting standards for financed emissions by ‘asset class’. It is based on the principle that financed emissions are a function of the supported entity’s emissions and an attribution factor based on the relationship between the outstanding amount loaned and the entity’s total value. Each transaction-level estimate therefore represents the proportion of the entity to which UKEF is a lender, guarantor, or insurance provider at a point in time.

UKEF offers three categories of products: direct lending, which is closest in nature to the project finance PCAF asset class; guarantees, for which there is no specific PCAF asset class guidance; and export credit insurance, for which there is no specific PCAF asset class guidance. Approximately 13% of UKEF’s financial exposure as at January 2024 is in direct lending products, with the remainder in guarantees and export credit insurance. Because methodologies for export credit insurance and guarantees are currently undeveloped in the PCAF Standard, and as an ECA with a portfolio materially weighted towards guarantees, we do not consider it appropriate to only attribute emissions from our guarantee exposure when the guarantee is called, as set out in the PCAF Standard.

Therefore, in the absence of a standardised approach among ECAs and a PCAF methodology for ECAs, UKEF has developed a double disclosure approach using two methodologies, informed by the PCAF guidance where relevant. Under this approach we estimate the financed emissions associated with our current portfolio exposure on both an ‘amount at risk’ (AAR) and ‘expected loss’ (EL) basis[2].

3. AAR methodology

AAR is equivalent to the accounting term “contingent liability”. This represents the unexpired portion of the total risks supported by UKEF, essentially amounts still owed to banks or exporters where UKEF could face a claim. AAR would normally be less than maximum liability by the amount of expired risk. It is the measurement of exposure for issued business.

The AAR methodology takes a “follow the money approach”, whereby financed emissions for all products are equal to: ∑ entity emissions × attribution factor, where attribution factor = AAR ÷ value of entity.

This approach does not make any distinction between loans, insurance products and guarantees. This approach helps us to understand the carbon intensity of transactions UKEF supports and provides a basis for managing the emissions associated with our support over time. It represents full ‘double counting’ of financed emissions between UKEF and other financiers on all guarantee or insurance facilities.

4. EL methodology

EL is the anticipated average loss over the relevant time horizon. For individual transactions, the statistical estimate of the most likely financial outcome on a transaction, based on the likelihood of default and estimates of recoveries.

The EL methodology takes a “follow the risk approach”, whereby financed emissions for insurance and guarantee products are equal to: ∑ entity emissions × attribution factor, where attribution factor = EL ÷ value of entity and financed emissions for direct lending products are equal to: ∑ entity emissions × attribution factor, where attribution factor = AAR ÷ value of entity.

The EL estimate is an attempt to better represent the economic role that export credit insurance and guarantees play in transactions. This approach is more consistent with our approach to understanding and managing the risk associated with our guarantee and insurance portfolio. This approach supports the reduction of double counting of emissions between financiers, which in aggregate should support a more accurate global picture of financed emissions.

As it is a risk-based approach, the resulting estimates are affected by changes in the risk profile and financial performance of a given transaction, so year-on-year changes in the estimate will be due to factors additional to changes in emissions. The effect of this metric will also vary between transactions and products. For example, where expected recoveries mean loss given default factors are low – as in much of UKEF’s aviation exposure – this will lower the attribution factor.

The relative weighting of financed emissions between sectors is different under the two attribution approaches. Because EL is relative to risk, these differences stem from the differences in the distribution of risk between transactions.

5. Methodological approach – emissions estimates and timings

Our financed emissions are calculated by taking a snapshot of our portfolio, as at the start of January each year. As we estimate our financed emissions on a calendar year basis these estimates are not consistent with the financial year statements referenced elsewhere in the Annual Report and Accounts. This inconsistency is due to the additional time needed to prepare the financed emissions estimates, within annual reporting constraints.

Our financed emissions estimates comprise our customers’ Scope 1 and 2 emissions. From 2024, we also estimate financed emissions of customers’ upstream Scope 3 emissions for all sectors.

Scope 1	Direct emissions from our customer’s own or controlled sources
Scope 2	Indirect emissions from the generation of energy the customer purchases
Scope 3	Indirect emissions from the customer’s upstream value chain only

This is a change in approach from our 2023 disclosure which accounted for Scope 3 emissions for the oil and gas sector, and mining and metals. Customers’ modelled Scope 3 emissions are upstream only, reflecting data availability and restrictions of the environmentally extended input-output modelling approach (see below).

6. Methodological approach – inflation adjustments and exchange rates

PCAF emissions factors are not updated every year, therefore may need to be adjusted to reflect currency inflation (between the PCAF Database emissions factor year and the current year) when using economic activity-based emissions factors. Country-specific inflation factors are calculated based on inflation data from the IMF, using annual average inflation rates for the base year (the PCAF Database emissions factor year).

UKEF’s financed emissions are calculated and reported in GBP. Where asset values are not available in GBP, local currencies are adjusted to GBP based on exchange rates used in our financial accounting systems. IMF exchange rates are used as a complement where internal exchange rates are not available.

7. Data quality – key issues

We use a combination of reported data, physical activity-based estimates and modelled data to estimate our financed emissions. In general, the highest data quality scoring source is preferred (see below on data quality scores). However, as emissions reporting is not yet required routinely across all economic activities in all geographies, there are significant gaps in the availability of emissions reporting in UKEF’s portfolio.

We rely on external sources for climate-related data. Key external sources include the PCAF Database for economic activity-based emissions factors, public reporting from companies where available, as well as non-public information shared including through environmental and social monitoring activity for projects supported.

The quality and accuracy of data varies widely between sources and between that available for individual transactions. Limited checks are performed to verify the accuracy of data sources; we are reliant on the accuracy of data sourced from customers and public sources. This is reflected partially in the data quality scoring, which includes verification in the hierarchy. Common issues include low resolution, consistency and transparency of company reported data, poor sectoral coverage and material time lags between emissions and financial data. Due to these issues, our financed emissions estimates are a high-level estimate of our customers’ activities for a given time period, using the data available at that point in time.

Using the highest data quality scoring source can result in time-series inconsistencies. For example, where reported emissions are not available for the current disclosure year, the most recent available reported or physical activity-based estimated emissions are used in preference to modelled emissions. The numerator and/or denominator of the attribution factor may also represent a different point in time to the emissions data. We do not currently have a time threshold for reverting to modelled emissions where there are time inconsistencies.

Data limitations mean it is necessary to use judgement in estimating financed emissions. Inaccuracies in our internal data also feed through into estimates. For example, there is known inconsistency in internal data between sector assignation, which impacts modelled estimates and disclosure of emissions by sector. This is corrected manually to a limited extent (<3% of individual transaction lines), but all individual transactions have not been checked for sector accuracy. Professional judgement is exercised in manually assigning sectors to transactions.

Additionally, where modelled estimates are used, the available emissions factors in many cases do not align directly with the specific economic activity supported. These modelled estimates are best understood as a high-level approximation of emissions associated with a broad economic activity, not an accurate representation of the specific activity supported.

Estimates are an attempt to best represent financed emissions associated with the actual activity that has taken place in a given year, within the prevalent constraints on data quality and availability. Where a transaction represents a project in construction, reported or physical activity-based estimated construction phase emissions data is used in preference. Where this data is not available, however, the transaction is not included in the estimate, as using modelled data for the operational phase would represent an entirely different category of activity which would materially overestimate the associated emissions that actually occurred in a given year. As at January 2024, 4% of UKEF’s total AAR (<1% EL) is transactions in construction phase where construction emissions are available and reported, 27% of UKEF’s total AAR (45% EL) is in transactions in construction phase where construction emissions are not available and therefore not reported.

Judgement is also exercised in selecting data to best represent the denominator of the attribution factor – the value of the entity. The PCAF Standard preferred metrics of Enterprise Value Including Cash, or Project Equity + Debt (where debt represents financing instruments that require repayment by the borrower) are used where these are available. EVIC values are obtained from S&P Capital IQ Pro. Where these are not used, however, professional judgement is exercised to select the best available data to represent the value of the entity. In some cases, this may be limited to the debt UKEF is party to, for example the contract value. Note EVIC values utilise April 2024 figures, other asset values utilise a variety of dates as best available, these are not fully consistent with the January 2024 outstanding exposure values.

8. Data quality - results

The PCAF Standard sets out a data quality framework, ranging from 1 to 5, with 1 being the highest data quality and 5 being the lowest. Each transaction is assigned a data quality score. The resulting exposure-weighted data quality score indicates at high level the data quality of the overall financed emissions estimate.

Our 2024 financed emissions estimates on an amount at risk basis have a weighted PCAF data quality score of approximately 3.06 (out of 5, where 5 is the poorest data quality) for Scopes 1 & 2, and 5 for Scope 3. On an expected loss basis, the PCAF data quality score is approximately 4.05 for Scopes 1 & 2, and 5 for Scope 3. These scores indicate our continued reliance on relatively low-quality data. (See box below for more information on limitations of the environmentally extended input-output modelling approach).

Environmentally extended input output modelling

• To fill gaps in emissions reporting, financial institutions can use environmentally-extended input-output models such as that provided by the PCAF Database, to derive sectoral emissions factor based estimates.

• This approach allows an institution to estimate, in broad terms, the financed emissions that arise from support provided to a given economic activity within a given geography. This approach is limited in its accuracy because it is based on abstractions which may not hold true for any given transaction.

• For example, input-output modelled estimates have a PCAF data quality score of 5, while reported and verified emissions have a score of 1.

• For more information on data quality see the PCAF Global Greenhouse Gas Accounting and Reporting Standard for the Financial Industry.

For UKEF’s aviation portfolio, all asset-based emissions are based on data from PACE[3]. PACE is a qualified data provider of the Pegasus Guidelines[4], a voluntary emissions measurement and reporting standard for the aviation sector. As and when PACE receive assurance on their emissions data in future periods UKEF expects a large increase of the portfolio’s PCAF data quality score, affecting the overall weighted PCAF data quality score.

Going forward we expect to continue to decrease reliance on environmentally extended input-output modelling as we collect more accurate transaction-specific emissions data from customers. We are reliant on the further development in emissions reporting practice to continue improving the data quality of UKEF’s estimates. As a PCAF signatory we will continue to work with industry peers and partners to enable further improvements in data quality and consistency of methodologies.

9. Restatement

As stated above, publishing estimates in the current environment of high uncertainty around both data and ECA financed emissions methodologies carries a high degree of risk. Restatement of the financed emissions baseline and/or years since the baseline will be necessary to help ensure the consistency and comparability of financed emissions estimates over time. The key areas of change which we would expect to prompt a baseline restatement are:

• Changes to the financed emissions methodologies used

• Errors such as a failure to carry out our methodology in line with the Basis of Reporting

• Changes to external data which drive a material change (>10%)

We expect our restatement policy to evolve with further industry guidance and as our approach to these estimates develops over time to reduce uncertainty.

10. Review and assurance

Financed emissions estimates are compiled internally and signed off by the Senior Civil Servant (Band 1) responsible for Climate Change and Sustainability, subject to the limited checks described above. For our 2024 estimate, we have obtained a limited assurance under the International Standard on Assurance Engagements (ISAE) 3000.

---

[1] Available at https://carbonaccountingfinancials.com/standard

[2] For definitions, see Glossary in UK Export Finance Annual Report and Accounts 2023/24

[3] Platform for Analysing Carbon Emissions - PACE (pace-esg.com)

[4] Pegasus Guidelines for the Aviation Sector - NDC Partnership

[i] PCAF makes no representation or warranties of any kind on the Database whether express, implied, statutory or other. This includes without limitation warranties of title, merchantability, fitness for a particular purpose, non-infringement, absence of latent or other defects, accuracy, or the presence or absence of errors, whether or not known or discoverable. PCAF disclaims any liability for any loss or damage resulting from use of information in this Database or reliance or decisions based upon the Outputs or Contents of the Database. Each Authorised User and/or third parties are advised that they are responsible for reliance on the database, data, information, findings and opinions provided by PCAF. Any reference to a specific data source, data provider or service by trade name, trademark, manufacturer, or otherwise, does not constitute or imply an endorsement, recommendation, or favouring by PCAF.