GENERAL COMMENTS CONCERNING THE OVERALL CONSULTATION:
The financial industry has a very important role in the transition of industry and society towards sustainability goals. Critical to success is the commitment of banks to actively encourage clients to make the transition and to provide the needed financing, the scale of which is substantial. With regard to prudential regulation it is key that banks manage the risks related to that transition pathway properly.
We expressly support a number of basic ideas in the discussion paper.
- We generally support the guiding principle of following a risk-based approach to the consideration of climate-related and environmental risks in the prudential framework.
- We support the view that ESG risks are not a new separate risk category but act as risk drivers or risk factors of familiar risk categories such as credit or market risk.
- We support the approach of first examining more closely the extent to which existing internal models can already, or could in future, cover ESG risks under Pillar I (or Pillar II). This is highly important in order to avoid double counting.
- In addition, we share the view of the French Council Presidency that there is not only a lack of consensus at international level but above all a lack of empirical data on the actual impact of ESG risks on Pillar I requirements, and that it would therefore be premature to adjust Pillar I rules at this stage.
We also consider it extremely important to conduct further discussion of the role of environmental risks in the prudential framework first and foremost at Basel Committee level. We would welcome it if the European representatives on the Basel Committee could put forward the positions of the German Banking Industry Committee (GBIC) for consideration in the course of further discussions. Global coordination is absolutely essential and European lawmakers should not pre-empt the debate by introducing their own measures.
COMMENTS CONCERNING QUESTION 1:
We agree with EBA’s risk-based approach to the consideration of environmental risk in the prudential framework and believe that the consideration of social risks should follow this risk-based approach as well. We therefore consider exposures associated with social objectives and/or subject to social impacts currently outside the scope of further Pillar I measures. No clear definitions have yet been agreed (EU social taxonomy, which is not a risk measurement tool) and there is a lack of (historical) data and methodologies for measuring social risk.
We acknowledge social risk as a topic for future integration into a holistic ESG target picture. Banks naturally already have many initiatives in place to support social aspects. The focus should initially be on integrating climate and environmental risks into the prudential framework.
Yes, we agree that liquidity will not be significantly affected by environmental risks.
Nevertheless, the impact needs to be regularly evaluated within the overall environmental risk materiality assessment as envisaged by the ECB Guide on climate-related and environmental risks under Pillar II.
We also fully agree that leverage ratios will not be significantly affected by environmental risks. Furthermore, the inclusion of risk-based aspects would be at odds with the fundamental concept of the leverage ratio as a non-risk-based ratio.
Environmental risks (transition as well as physical) are expected to materialise over a long time horizon and thus do not require a recalibration of the one-year perspective in the Pillar I approach.
As also highlighted by the Bank of England in the results of the Climate Biennial Exploratory Scenario, “the overall costs to these firms from the transition to net zero should be bearable without substantial impacts on firms’ capital positions – for example through a combination of lower retained earnings and increases in lending rates to sectors where risks increase, and also because not all of the losses on insurers’ investments would ultimately fall on shareholders”. We therefore agree with the conclusion of the Bank of England that “regulatory capital is not an appropriate tool to address the underlying causes of climate change (ie greenhouse gas emissions across the economy). It would likely be both less effective than other possible direct climate policy interventions and could potentially give rise to unintended consequences for firms’ safety and soundness. The responsibility for addressing the causes of climate change ultimately lies with governments, businesses and households.”
We agree with the risk-based approach of EBA’s discussion paper and also agree that (historical) data and evidence are not yet available on a scale that properly allows an inclusion in the Pillar I framework. But even if the situation in relation to historical data improves, there are reasonable grounds for doubting whether – given the future-related nature of climate risks – such data could offer a calibration basis for their inclusion in Pillar I.
Environmental risks can be covered in Pillar II by II via discounting of future cash flows. As Pillar I is already “informed” by Pillar II via existing ICAAP (e.g. scenario simulations, discounted P&L impacts from Pillar II are included risk by risk in Pillar I), no further tools are needed. However, we do not believe that internal capital for long-term environmental risks is justified or necessary.
In addition to the risk-sensitive perspective, environmental risks are certainly not about overall risks but about the exposure of the individual bank portfolio. Hence, environmental risks should be assessed on a bank-specific basis pursuing a more granular approach in line with the bank’s portfolios, business model and strategy.
Finally, an additional risk category would reframe existing risk management practices as an additional risk assessment tailored to environmental risks would need to be pursued. Additional capital requirements may result in consequence of the additional risk category even if there is no change to the credit risk assessment. We therefore support the EBA’s view that ESG risks are not a new, separate risk category but act as risk drivers or risk factors of familiar risk categories such as credit or market risk. As stated above, additional capital requirements for long-term environmental risks would not serve a meaningful purpose.
We strongly believe that while consideration of the “double materiality” principle is important for banks’ steering, prudential regulation should focus solely on incurred risk in line with the EBA DP’s risk-based approach, which we support.
We would like to emphasise the need to avoid double counting of risks. In addition, if a parallel concept to that for financial reporting is envisaged, we would like to refer to our answer below on operational risks such as back-feeding reputational risks: these should be kept in Pillar II assessments and should not interfere with Pillar I quantitative requirements.
A key issue is the lack of European harmonisation of (new) data requirements and the regulation of both implementation timelines and data definitions.
Implementation timelines: Several banking regulations and supervisory exercises (e.g. CRR disclosure, ECB climate risk stress test) impose strict requirements on data related to environmental risk while the relevant data disclosure regulations for corporates (e.g. CSRD, Article 8 of the Taxonomy Regulation) are yet to be implemented. While the aim should be to harmonise future regulations, it is important that banks are allowed to use prudent and reliable proxy methods to calculate data that are not (yet) available.
Data definitions (e.g. EPC label classification, scope 3 GHG measurement) are not yet fully consistent across SSM countries. The aim should be to harmonise the most important data types. The data situation could also be vastly improved by creating a harmonised and publicly accessible EPC register at least at European level, along the lines of that envisaged for the future European Single Access Point (ESAP).
The disclosure of exposures to the “top 20 polluters” as envisaged by the CRR will be especially difficult for banks without an agreed list of relevant corporates. A central register/list should be provided by regulators.
We consider a risk-based approach to environmental risks to be absolutely essential.
Banks must assess default risks appropriately to maintain the financial stability of the bank, the lending market, the banking sector and the financial system as a whole.
In general, we agree with the risk-based approach of EBA’s discussion paper and also agree that (historical) data and evidence are not yet available on a scale that would allow the appropriate inclusion of environmental risks in the Pillar I framework.
Moreover, we would like to stress that there should be no dependency on the taxonomy for risk management purposes as this is not a risk measurement framework.
It is vital not to alter the current timeframe as the whole risk assessment and risk management process as well as investment decisions and regulators’ decisions on capital requirements are framed by the current time horizon.
Any change to the current time horizon would have multiple unintended and incalculable consequences. For example, current capital ratios would change, resulting in completely incomparable ratios and data.
Environmental risks (transition as well as physical) are expected to materialise over a long time horizon, which can be covered in Pillar II (by discounting future cash flows) and do not require recalibration of the one-year Pillar I approach.
Pillar I is not the appropriate tool for reflecting environmental risks because the impact will be on micro level, not on macro level. Hence, no industry-wide measures are required but bank-specific analysis needs to be pursued, e.g. by stress testing and scenario analysis.
Many institutions have performed scenario-based sensitivity and stress analysis for transition and physical risk. The scenario analyses showed that, for many banks, transition risk is more material than physical risk for corporates and that climate risks generally have a moderate impact on portfolio level, with a higher impact in certain sectors depending on the scenario.
We would like to highlight that the regulatory requirements regarding the statistical significance and informative value of a new (ESG-related) risk driver in internal rating models are very high and cannot be met at the moment.
The main challenges currently faced by lenders and rating agencies relate to the availability and comparability of environmental/ESG disclosures by clients (both within and across industries) and the difficulty of translating/incorporating such varied (and still rather patchy) data and information into the wider business and financial risk assessment that forms the basis of every credit rating. Data sets for rating development might be insufficient to prove the significance of environmental risk factors (e.g. due to the short time series of ESG-related variables).
With respect to models, it is not clear whether a defaulted company has defaulted due to an environmental-risk-associated event.
Incentives: Banks may face a reputational risk if the method is not properly validated or if the market/authorities expect incorporation of ESG factors even if there is no statistical evidence.
It should also be borne in mind that banks depend on the cooperation of rated companies, which have no incentive to report poor ESG-related KPIs.
Rating agencies’ reports on a given client or sector – where available – are part of the information gathered and considered as part of the internal risk assessment in some banks. Others make no use of external ratings.
To avoid imposing a substantial operational burden or complexity, broadened due diligence requirements for the integration of climate and environmental risks should be implemented with a clear focus on portfolios/asset classes where this is most relevant. So the “proportionality principle” mentioned in the discussion paper could be applied in a way that enables a bank to use its judgement to identify the portfolios for which these additional due diligence requirements are relevant and useful (the definition could be based on the size of asset, sectors, products).
The usage of external data sources (where available) should be considered as an alternative source for gathering the necessary additional data/information.
Client due diligence is already adapting to the need to capture climate and environmental risk aspects. This reflects both internal risk management considerations as well as debt and equity investors’ increased awareness and sensitivity to this risk dimension.
The gradual increase in client climate and environmental related disclosures (be they government mandated or voluntary) – together with the trend towards greater standardisation and comparability of disclosed metrics/KPIs – is expected to increase the consistency of climate and environmental risk assessments as well as the resulting insight gained.
In many cases, environmental data are not available from the client, central databases or external data providers. Regulatory and/or market standards for data requirements are fast developing. Yet private clients, in particular, have neither the knowledge nor the financial capacity to provide audited documentation on environmental risks (associated with their real estate, for example). In addition, there are currently not enough experts in the market to cover all these data requests (e.g. to verify EPC labels or calculate the energy supply from solar panels).
Currently no. PD, LGD and “first of all” collateral valuations – as applicable to different types of exposures – remain the most appropriate analytical tools for consolidating the risk assessment of a counterparty, including its exposure to C&E risk. However, these parameters need to be continuously reviewed in order to incorporate special cases of environmental risks.
In our view, the CRR3 proposal’s clarification of energy efficiency improvements brings enough risk sensitiveness to the framework for exposures secured by immovable properties since increasing physical risks should typically lead to decreased market value whereas mitigating transition risk by energy efficiency improvements should be taken into account in the determined market value.
The EU taxonomy requires energy efficiency to be demonstrated by EPC documentation. For EPCs, there is a huge data gap as well as a lack of experts capable of providing sufficient documentation for transactions already on the balance sheet (see also our reply to Q11). For data availability, a central European governmental initiative is needed to collect standardised data (at least within Europe). Until such a central, (at least) Europe-wide EPC register is established, which will require some considerable time, well-founded proxy approaches for deriving EPC ratings based on a few relevant input factors should continue to be accepted by supervisory authorities (for lack of an alternative).
The challenges are even greater for energy efficiency improvements – be they in the form of retrofitting or only replacing a heating system. Replacing a 30-year-old heating system, for example, will increase the efficiency from 60% to 90%, which will lead to a significant saving of energy consumption. However, a marginal investment of €8,000 to €12,000 by a private household would give rise to €100 to €300 in additional costs to generate a new EPC. A simpler, cheaper EPC would need at least one year, better three years of energy consumption of the new heating system. Larger renovation projects for CRE transactions face a similar challenge since it is hard to prove the effect of increased energy efficiency at the time of origination of the loan. As a result, the risk weight will only decrease at a later stage.
In order to remedy this, loans for renovating real estate should already be assigned a lower risk weight at the time of origination of the loan. To simplify the process, an additional budget for renovation should receive a low risk weight and have low documentation requirements up to a certain percentage of the property value depending on the construction year of the property. For private households, the risk weight for an additional loan for renovation should be zero up to €50,000 if, and only if, the real estate is obtained as collateral by the bank.
Like the Infrastructure Support Factor (ISF), the definition of “high-quality” project finance is already very restrictive; adding environmental criteria would basically align this concept with the ISF criteria, leading to the same usability/applicability issues.
In addition to that we understand that CRR proposes a lower risk weight for high-quality project finance/object finance. This lower risk weight should reflect the better risk profile of such assets.
With this in mind, we do not support the explicit consideration of environmental criteria for high-quality project finance and high-quality object finance exposures introduced in CRR3 as it would run counter to the overall risk-based approach to a certain extent. Environmental aspects should only be considered to the extent that they are relevant for the risk assessment.
Given current data availability, further risk differentiation would not be justified (see our reply to Q11). If additional data are provided, these could be further analysed in order to see whether risk differentiation would be appropriate.
Since environmental risks are more likely to materialise in the long-term, but typically not within the next year, environmental risk drivers should only be incorporated into IRB models if they significantly improve the models statistically. The most appropriate reflection would be in the models (corporates -> PD; real estate -> LGD), but there is currently a lack of evidence.
If environmental risks materialise for a single counterparty, this will be reflected by a deterioration of scores that are typically already part of internal rating models, such as scores for the qualitative risk assessment, target balance sheets and credit spreads.
At the moment we see no necessity to change corresponding rating scorecards.
For real estate, energy efficiency should be used to reduce risk weights for mortgages in the F-IRB approach in order to support higher investments in, and for, energy efficient buildings.
It is not possible to give a general answer here. However, some banks reported that they do not explicitly incorporate ESG risk factors into their IRB models. If environmental risks had a short-term impact on PD, it could be captured in target financial KPIs in the finance or market modules of IRB models.
Many banks plan to assess whether ESG factors can improve PD and LGD models. At the moment, many banks integrate climate risk scores into their credit process, but not quantitatively into PD/LGD models.
The scope is defined in such a narrow way that the ISF has become unusable. It could be strengthened by making it effectively feasible to apply, meaning that the scope needs to be reframed in a broader and clearer way.
It is conceivable that sustainable financing has lower ESG risks overall than conventional loans. However, in the absence of clear evidence at the moment, neither a general Pillar I requirement nor a sector-wide risk adjustment factor would be appropriate from a risk perspective. Credit risk should assess the default risk of an exposure class. As long as there is no clear evidence of increased default risks of exposures affected by environmental risks, there should not be any additional quantitative Pillar 1 requirements. Should significantly lower or higher risks become apparent within the one-year horizon considered for sustainable or conventional financing, we expect this will be reflected in well calibrated IRB approach processes. For individual assessments, instruments are already taken into consideration in Pillar II. This should suffice for the time being.
We do not support adjustment factors to tackle environmental risks. The risk of double counting is clear – as a result not only of a potential adjustment factor but also of existing prudential requirements such as the systemic risk buffer, capital conversation buffer or Pillar II capital add-ons. Therefore, regulators need to explore how this double counting can be avoided.
First of all, it should be borne in mind that the FRTB standardised approach has to be applied in parallel with the FRTB IMA and serves as a fallback solution in the event that models are deemed of poor quality, for example. For this reason, the treatment of environmental risks in the FRTB standardised approach and FRTB IMA should generally be consistent.
Secondly, the current standardised approach will continue to be used (as a simplified standardised approach) in the EU in addition to the two approaches addressed in the discussion paper. If recalibration really proves necessary due to environmental risks, this will probably also apply to the simplified approach, so adjustment options need to be analysed in this area as well. We are firmly opposed to the idea of increasing risk weights in the simplified standardised approach at the level of market risk categories regardless of the actual environmental risks in a bank’s portfolio, as has already happened in the FRTB context. Here, too, the Basel Committee should possibly explore how to reflect environmental risks in a differentiated manner.
For various reasons, we oppose the idea of adding surcharges to risk weights based on forward-looking scenarios. Such an approach would be at odds with the existing Basel Pillar 1 approach of the FRTB standardised approach and would be based on highly uncertain and subjective assessments of scenarios, whose effects on banks could vary widely depending on their individual market risk positions.
The introduction of additional scenario considerations would be a suitable risk management instrument, in our view, but this should be subject to Pillar 2only. The more detailed derivation of risk weight allocations as a result of adding certain factors relevant to environmental risk (such as the dependence of the risk factor for equities on the economic activity and sector, for example) would need to be clearly empirical in nature but would remain an extremely broad-brush approach since the individual bank’s specific market risk exposures would not be reflected. Introducing a new risk factor specifically for environmental risks would also need to be subject to empirical evidence of the effects of environmental risks on relevant trading or banking book instruments. As a general principle, in-depth analysis at Basel level is essential to properly investigate the effects on capital requirements and the exact calibration of capital approaches.
Regarding the basic freedom to choose which method to apply when using internal models, the changes introduced under the Fundamental Review of the Trading Book were made in part as a result of greater restrictions on discretionary leeway. Given the increasing subjectivity and possible double counting as a result of the difficulty of separating out the specific effects of environmental risks from general volatility, we have serious reservations about including environmental risks within models by adjusting observed historical data to accommodate possible future dynamics. This also applies to the possible mapping outside models by adjusting the consideration of event risks. Given that empirical observations and model developments are still in their infancy, we would prefer more experience to be gained on capturing these risks in Pillar II before making any adjustments to the Pillar I methodology of capital approaches (and if such adjustments are made, then initially only at Basel level). This approach could enable a ranges of practices to be established that could feed into discussions about future adjustments to Pillar I.
Banks with significant market risk exposures should be given the opportunity under Pillar II to gradually refine their models when a larger empirical database becomes available before any decision is made on adjusting Pillar I.
First of all, we consider the existing event types to be sufficient as they cover all sorts of events that could be caused by climate risks. In order to be able to identify losses from environmental risks, a flag in addition to the existing event types could be introduced as proposed in the paper.
The main issue is the lack of an official universal taxonomy or categorisation of ESG risks defined by supervisors along the lines of Basel event type categorisation (including a level 3 example such as EBA guidelines). Likewise, the EBA should give guidance on assigning loss data to the ESG risk categories for each Basel event type based on data available to the EBA.
The current classification by event type focuses on the effect of operational risk (e.g. physical damage). One way of increasing transparency would be to add a cause dimension (e.g. weather effects). A challenge with respect to climate risk will be to differentiate between normal weather events and climate risk (e.g. rainfall as a natural normal event and rainfall as a climate risk event).
We see another challenge with respect to physical risks: as pointed out in the paper, access to climate data is limited, as are forecast possibilities. Where banks are not able to access or model these, risk assessments by regulators could be helpful, such as in the form of risk scores for certain regions in order to assess the physical risk to a bank’s own infrastructure or third-party providers in these regions.
As correctly summarised in the EBA’s paper, the Basel IV standardised measurement approach (SMA) is not designed to give this perspective. However, this deficiency applies to all OR risks, not just environmentally driven ones. Given the progress of Basel IV implementation, tweaking the SMA is not a realistic scenario, so any specific treatment of ESG risks should become/remain part of the Pillar 2 framework.
Data sources for ESG “forecast” data (expected loss development from ESG risk) recommended by regulatory authorities would be appreciated. Due to the lack of regulatory requirements and guidelines, it is not possible to assess such expected losses as things stand. Moreover, bank-specific activities diverge and hence results are not comparable (in the same way as at the beginning of the advanced measurement approach for operational risk).
In the Pillar II OpRisk framework, scenario analysis of risk severity and probability based on expert assessments can provide a forward-looking perspective which can be taken into account for internal capital adequacy purposes.
There were good reasons for excluding strategic and reputational risks from the definition of operational risk and from Pillar 1. ESG risk drivers could nevertheless have a significant impact on these risk types. No changes should be made to how Pillar 1 functions, however, and these risk factors should be dealt with solely under Pillar 2, and thus in the SREP, since no generally applicable capital approaches are available to enable risk-weighted assets to be determined.
It is essential that any regulatory framework treats ESG factors as risk drivers (i.e. similar to conduct risks) rather than as a separate new risk type. Otherwise, banks will face serious challenges in rearranging risk taxonomies with many unintended consequences for risk management and reporting processes.
Additionally, it is very important to highlight the necessity of an overarching approach to ESG risk management that considers all relevant second lines (compliance, risk, legal, business continuity and so on) and to issue correspondingly holistic guidelines to create awareness that all these units need to get involved.
As explained in our reply to Q27, expert-based scenario analysis is a possible way of obtaining meaningful insights in situations where historical data are scarce, such as on a new/developing risk like climate risk.
We support EBA’s opinion that new concentration limits might have a negative impact on financing a counterparty’s transitioning to environmentally sustainable activities. Against this background, we consider the reporting and monitoring of potential concentration risk – complemented by Pillar II measures if necessary – more effective. Tailor-made supervisory responses based on meaningful/powerful reporting will prevent adverse effects on transitioning. We therefore also agree with the EBA that expanding the existing large exposure framework is not the correct path for addressing this perceived risk due to its focus on groups of connected clients.
From the relevant section of the discussion paper, we understand that the potential thrust of such concentration risks is not limited to a high risk of natural catastrophes like flooding, earthquakes or wildfires, but could also potentially include steering exposure concentrations of exposure to environmentally harmful sectors, like the mentioned carbon-intensive industries. The latter, in particular, should not be subject to financial regulation, since this would be an attempt to enforce non-financial goals through financial regulation.
Environmental risk should be defined as a risk driver of all relevant risk types in Pillar II/ICAAP to ensure the detection of inter-risk concentrations as defined in the ECB’s Guide on climate related and environmental risks (“Consequently, physical and transition risks are drivers of existing risk, in particular credit risk, operational risk, market risk and liquidity risk, as well as non-Pillar 1 risks such as migration risk, credit spread risk in the banking book, real estate risk and strategic risk. Climate-related and environmental risks may, in fact, be drivers of several different risk categories and sub-categories of existing risk categories simultaneously.”).
The simulation of consistent environmental scenarios including all relevant risk types should be conducted to clearly identify overall environmental risks as envisaged in the ECB Guide as well (“In line with the ECB Guide to the ICAAP, institutions are expected to consider in their forward-looking capital adequacy assessment any risks, and any concentration within and between those risks, that may arise from relevant changes in their operating environment.”). Currently, no relevant environmental concentration risk can be observed (at least in the short term).
Regarding current/future reporting requirements which lend themselves to effective monitoring of environment-related concentration risk, we would like to draw attention to other reporting requirements that institutions are obliged to comply with:
- Current LEX reporting: Article 394(1), sentence 3 of the CRR requires institutions to report – as part of LEX reporting – all exposures of a value greater or equal to €300m (but less than 10% of the institution’s Tier 1 capital). The reporting (template LE1/C27) comprises the residence and the sector (plus NACE code) of the counterparty. For large institutions, a threshold of €300m makes it necessary to report all significant counterparties.
- Current Pillar 3 disclosure and future ESG reporting: templates 2 and 3 of EBA/ITS/2022/01 on prudential disclosures on ESG risks in accordance with Article 449a of the CRR already makes a distinction between “exposures towards sectors that highly contribute to climate change” and “exposures towards sectors other than those that highly contribute to climate change”. Besides this, a detailed granular sectoral breakdown is required for all banking book exposures classified as “towards sectors that highly contribute to climate change”.
- The disclosure templates mentioned could be integrated into future ESG reporting requirements (in accordance with Article 430(1)h of the CRR3 proposal) without requiring any significant extra effort by institutions. We therefore see no need for additional sectoral concentration reporting. If considered necessary, a similar geographical reporting could be integrated into future ESG reporting.
Overall, we consider the current reporting requirements, including those intended in the CRR3 proposal, sufficient to cover the need for prudential information on sectoral and geographical breakdowns. As regards counterparty concentration risk, a reporting threshold of €300m is sufficient to cover all significant counterparty risks of large institutions. Before planning new concentration reporting requirements, the appropriateness of current (or already intended) reporting requirements should be reviewed. If additional reporting requirements are considered necessary, they should be harmonised with the existing requirements to avoid any duplicative reporting requirements. To avoid an ongoing expansion of reporting requirements – and a corresponding reporting burden on institutions – the possibility of dropping certain existing reporting requirements should be examined.
As we see it, regulatory limits have been expanded significantly over the past decades and the Basel regime has not even been transposed into European legislation yet. Adding another regulatory limit or additional rules would overburden the sector.
There is currently no evidence of bank defaults caused by overexposure to natural catastrophes. (And even if there were, this might primarily be an issue for insurers or reinsurers.) It is in the interest of banks to build a balanced credit and asset portfolio which takes concentration risks of all kinds into account through regular reviews of industry-specific portfolios and their idiosyncrasies, for example.
The potential cost of introducing such a limit, i.e. building the methodologies and IT systems at each bank to monitor it in accordance with the (yet to be established) regulatory rules, and demonstrating and auditing compliance for the entire EU banking industry risks exceeding the potential economic benefit.