Response to discussion paper on the role of environmental risk in the prudential framework
Go back
In a nutshell, AIFIRM deems that the creation of a sub-bucket incorporating the environmental risk is the best way forward, among all the approaches presented.
New Risk Factor
As clearly pointed out in the EBA paper: “while environmental risks may not lead to the introduction of new risk factors per se, they may affect the magnitude of their shocks”. Building on this statement shared by the EBA in the DP, AIFIRM assesses it’s very difficult to disentangle the environmental Risk component embedded into market prices.
As a consequence, irrespective of the framework adopted within the standardized approach to provide a capital requirement for such risk, it would result very difficult, or even impossible to identify the appropriate financial instruments or calibrate the shocks/stresses to observable market prices.
Besides the evident difficulties in creating an environmentally related risk factor ex novo, a comprehensive framework would need to be established in order to assess which positions should fall into this risk class, and consequently be excluded from the other existing risk classes. The possible choice of this approach would need to be gauged to avoid double-counting.
New Bucket
We believe environmental risk to be a peculiar driver of risk, which spills over into the other risk classes. As such, defining a specific independent environmentally related bucket would not be feasible.
Instead, we believe that the introduction of environmentally related sub-buckets into the existing bucketing process for each risk class might be the most effective and feasible path among the ones proposed so far.
One proposal could be structured around the implementation of three (or more) sub-buckets: for example, one sub-bucket encompassing financial assets carrying high environmental risk, one comprising assets subject to environmental risk only in a limited way and one grouping assets carrying low environmental risk. Each of these buckets could have different risk weights, the highest RW could be associated with the high environmental risk bucket, conversely the lowest RW could be associated with the low environmental risk bucket.
We deem helpful establishing a framework for setting environmental buckets for each risk class:
• For credit instruments further buckets can be introduced beyond the existing ones, after the “Sector” attribute. For each sector there can be sub-buckets subdividing environmental high performers and environmental low performers. For instance, a bond of an energy-producer with an investment-grade credit rating (which would, according to the CRR2, be allocated to bucket 5) could face a lower capital charge if the bond is finalized to financing a sustainable program such as energy production to be generated with renewable sources instead of being generated with fossil fuels. This would be possible by introducing further buckets accounting for environmental risk.
A lower capital charge for low environmental risk assets would also be consistent with empirical evidence found across the financial market, where it has been noted that, all else being equal, highly environmentally rated bonds have a slightly lower spread compared to their low rated counterparts (Polbennikov, Simon, et al. "ESG ratings and performance of corporate bonds." The Journal of Fixed Income 26.1 (2016): 21-41.)
For sovereign positions, the repartition between low environmental risk and high environmental risk could follow the natural progression of the sovereign debt instruments market. As now countries have started to differentiate between bonds issued to finance green projects and others, the use of sub-buckets might follow the exact same path. The European Union is embarking into an ambitious Net Zero project aimed at achieving zero carbon emission by 2050 (FitFor2050), financial assets issued by Member States specifically to bring this transformation forward could also be allocated to an environmental high-performer bucket.
• For equity further buckets can be introduced beyond the existing ones, after the “Sector” attribute. For each sector there can be different buckets subdividing environmental high-performers and environmental low-performers. The same example as above, specifically geared for equity might apply.
• For commodities, also in light of the ever-growing adoption of low-carbon fuel sources, it is possible to introduce this environmental risk bucket for all of the current commodity buckets, in order to foster the adoption of environmental risk for the commodities.
• For interest rate and forex instruments the application of an environmental bucket in the bucketing process is not evident, in AIFIRM’s opinion, positions in the scope of these two risk classes should be not assigned to any environmental-risk sub-buckets.
Nonetheless, the criteria to define what is considered to be environmental high and low performers needs to be clearly developed by regulatory authorities in a consistent and univocal approach which could then be used by financial institutions in assessing instruments in their trading books. This is the only way to provide comparable results among EU banks at least.
Increased risk weights
In our view this represents the simplest among the solutions proposed as increasing risk weights to account for environmental risk could lead to a more prudential requirement.
Nonetheless, there are some obstacles to be overcome in implementing this solution.
Firstly, a straightforward increase of the risk weights, without the differentiation between the assets that have low environmental risk and those that have high environmental risk, would result in a “flat-out” increase in the overall capital requirement. This approach would not take into account the comparative lower riskiness of environmental-virtuous assets in the Trading Book.
It is must also be noted that risk weights already partially incorporate environmental risk. Since only partial incorporation of risk is assumed, AIFIRM deems more reasonable to rely on the previous approach of including a further dimension to the bucketing process: such approach would allow for a marginal gearing of the existing weights in order to incorporate the environmental risk according to the effective exposure to environmental risk. With a flat increase of risk weights, an environmentally risk-adverse portfolio would face the same capital charge as a portfolio with assets comparatively more subject to environmental risk, provided that the nature and the direction of the two portfolios are equal.
Secondly, this increase in risk weights would most likely have some unintended consequences, such as an arbitrary increase in the capital charge faced by most banks for the positions in their trading book. Also, it would provide no incentives for them to switch their investment policies in favour of environmentally friendly alternatives.
Thirdly, the calibration of risk weights would be rather difficult to attain without a proper framework and the estimation of such an increase would rely on arbitrary data as environmental risk has not been properly monitored for long periods of time. The calibration of risk weights would need also regulatory inputs and the development of a new set of rules on behalf of the regulator.
RRAO
The discussion paper also mentions RRAO as a possible way of accounting for environmentally related risks. This framework is intended to address risks of complex products or products not fully covered by the SBA and default risk frameworks, but which are already identified by the risk management. Although being identifiable, environmental risk is present in plain-vanilla products and it is difficult to isolate depending on risk factors and asset class. In addition, such an approach will most probably lead to a strong divergence of Market Risk Department and Front Office’s pricing which is against the most recent direction the FRTB is leading us through (i.e., consistency between front office and risk management unit prices).
Concluding remarks for answer 23
To conclude, the approach we consider the most feasible is the elaboration of sub-buckets properly set for environmental risk. Still, we envisage two hurdles which must be overcome in order to set this method into motion. Firstly, sub-bucketing should not lead to an arbitrary increase of the capital charge for banks. Secondly, the regulator should provide financial institutions with proper risk weights, based on reliable data, although AIFIRM realizes how difficult their calibration might be.
However, a major limitation of modelling environmental risk within the IMA is that, as of now, there is no broad-based consensus on how environmental risk might be specifically isolated in either VaR or ES, thereby rendering its direct observation challenging.
As already mentioned in answer to question 23, at the moment AIFIRM does not deem accurate to define a new environmental risk factor. Risk Management may model curves, surfaces, vertices for each asset classes, but it may not unbundle a common risk factor like the environmental factor without seriously affecting the PLA.
Otherwise, considering environmental risk as a non-modellable risk factor seems not to be in line with the original aim of the non-modellable risk factor approach which should serve complex products only. As a result, such approach would cause an unintended penalization of all instruments with environmental intrinsic factors.
Another consideration AFIRM raised is related to the input to the models such as PDs or historical series of prices. They might be adjusted to estimate environmental risk but there are limitations. Firstly, the current regulatory framework does not allow to adjust historical data. Secondly, environmental risk and green assets are only now starting to make their way into the portfolio of investors in a meaningful amount; therefore, historical data is patchy and unreliable making the job of building parameters specifically geared to account for this type of risk extremely challenging.
The discussion paper also suggests incorporating a specific Add-On into the internal model. However, as the discussion paper is scant on details on how this would be implemented, it is difficult to assess it. This difficulty is further compounded by the recurring theme of the unreliability of environmental data and the divergence in how risks other than the ones normally considered are accounted for.
Instead of a change in the IMA regulatory framework, AIFIRM suggests leveraging the stress testing exercise, just like the one recently set up by the ECB.
AIFIRM proposes to consider the inclusion of environmental risk in the Pillar II framework instead of Pillar I to allow banks to capture the institution-specific nature of environmental risks.
In case changes to SA and IMA are applied, AIFIRM proposes to consider a two-way approach:
- keeping a framework as much as possible in line with the one currently in use/design (eventually with a slightly more prudent RWs) and
- designing a framework to provide incentives, by means of capital benefit, for banks willing to embark in a reassessment of their current framework (and having an environmentally safe portfolio).
All in all, one incentive tool that we estimate to be of relevance is the relative lower charge faced by environmentally risk proofed assets (as already mentioned in answering Q23). This point, along with the possibility of maintaining the current prudential framework, modified introducing new environmentally related adjustments, would be an incentive for financial institutions to rely on the new approach (when and if possible).
Q23: What are your views on possible approaches to incorporating environmental risks into the FRTB Standardised Approach? In particular, what are your views with respect to the various options presented: increase of the risk-weight, inclusion of an ESG component in the identification of the appropriate bucket, a new risk factor, and usage of the RRAO framework?
In the context of the Standardized Approach for Market Risk, the discussion paper structures the analysis around the latest regulatory development, the CRR2 and the CRR. By leveraging on the prudential regulatory framework, the EBA has put forward four main proposals regarding how it might be appropriate to incorporate environmental risk into the existing legislation, limiting, when possible, the creation of a new standard to account for this type of risk. AIFIRM discussed the proposed solutions by firstly analysing the introduction of the environmental risk factor which AIFIRM doesn’t see as an appropriate solution. Then AFIRM proceeds to analyse changes to the bucketing approach which is, in the view of AIFIRM, the best way forward (if considered as a sub-bucketing approach). Finally, AIFIRM’s market risk commission analyses the increase of risk weight and the application of RRAO: both, despite requiring less effort to be incorporated in the current regulation, seem to be less reasonable and less accurate solutions.In a nutshell, AIFIRM deems that the creation of a sub-bucket incorporating the environmental risk is the best way forward, among all the approaches presented.
New Risk Factor
As clearly pointed out in the EBA paper: “while environmental risks may not lead to the introduction of new risk factors per se, they may affect the magnitude of their shocks”. Building on this statement shared by the EBA in the DP, AIFIRM assesses it’s very difficult to disentangle the environmental Risk component embedded into market prices.
As a consequence, irrespective of the framework adopted within the standardized approach to provide a capital requirement for such risk, it would result very difficult, or even impossible to identify the appropriate financial instruments or calibrate the shocks/stresses to observable market prices.
Besides the evident difficulties in creating an environmentally related risk factor ex novo, a comprehensive framework would need to be established in order to assess which positions should fall into this risk class, and consequently be excluded from the other existing risk classes. The possible choice of this approach would need to be gauged to avoid double-counting.
New Bucket
We believe environmental risk to be a peculiar driver of risk, which spills over into the other risk classes. As such, defining a specific independent environmentally related bucket would not be feasible.
Instead, we believe that the introduction of environmentally related sub-buckets into the existing bucketing process for each risk class might be the most effective and feasible path among the ones proposed so far.
One proposal could be structured around the implementation of three (or more) sub-buckets: for example, one sub-bucket encompassing financial assets carrying high environmental risk, one comprising assets subject to environmental risk only in a limited way and one grouping assets carrying low environmental risk. Each of these buckets could have different risk weights, the highest RW could be associated with the high environmental risk bucket, conversely the lowest RW could be associated with the low environmental risk bucket.
We deem helpful establishing a framework for setting environmental buckets for each risk class:
• For credit instruments further buckets can be introduced beyond the existing ones, after the “Sector” attribute. For each sector there can be sub-buckets subdividing environmental high performers and environmental low performers. For instance, a bond of an energy-producer with an investment-grade credit rating (which would, according to the CRR2, be allocated to bucket 5) could face a lower capital charge if the bond is finalized to financing a sustainable program such as energy production to be generated with renewable sources instead of being generated with fossil fuels. This would be possible by introducing further buckets accounting for environmental risk.
A lower capital charge for low environmental risk assets would also be consistent with empirical evidence found across the financial market, where it has been noted that, all else being equal, highly environmentally rated bonds have a slightly lower spread compared to their low rated counterparts (Polbennikov, Simon, et al. "ESG ratings and performance of corporate bonds." The Journal of Fixed Income 26.1 (2016): 21-41.)
For sovereign positions, the repartition between low environmental risk and high environmental risk could follow the natural progression of the sovereign debt instruments market. As now countries have started to differentiate between bonds issued to finance green projects and others, the use of sub-buckets might follow the exact same path. The European Union is embarking into an ambitious Net Zero project aimed at achieving zero carbon emission by 2050 (FitFor2050), financial assets issued by Member States specifically to bring this transformation forward could also be allocated to an environmental high-performer bucket.
• For equity further buckets can be introduced beyond the existing ones, after the “Sector” attribute. For each sector there can be different buckets subdividing environmental high-performers and environmental low-performers. The same example as above, specifically geared for equity might apply.
• For commodities, also in light of the ever-growing adoption of low-carbon fuel sources, it is possible to introduce this environmental risk bucket for all of the current commodity buckets, in order to foster the adoption of environmental risk for the commodities.
• For interest rate and forex instruments the application of an environmental bucket in the bucketing process is not evident, in AIFIRM’s opinion, positions in the scope of these two risk classes should be not assigned to any environmental-risk sub-buckets.
Nonetheless, the criteria to define what is considered to be environmental high and low performers needs to be clearly developed by regulatory authorities in a consistent and univocal approach which could then be used by financial institutions in assessing instruments in their trading books. This is the only way to provide comparable results among EU banks at least.
Increased risk weights
In our view this represents the simplest among the solutions proposed as increasing risk weights to account for environmental risk could lead to a more prudential requirement.
Nonetheless, there are some obstacles to be overcome in implementing this solution.
Firstly, a straightforward increase of the risk weights, without the differentiation between the assets that have low environmental risk and those that have high environmental risk, would result in a “flat-out” increase in the overall capital requirement. This approach would not take into account the comparative lower riskiness of environmental-virtuous assets in the Trading Book.
It is must also be noted that risk weights already partially incorporate environmental risk. Since only partial incorporation of risk is assumed, AIFIRM deems more reasonable to rely on the previous approach of including a further dimension to the bucketing process: such approach would allow for a marginal gearing of the existing weights in order to incorporate the environmental risk according to the effective exposure to environmental risk. With a flat increase of risk weights, an environmentally risk-adverse portfolio would face the same capital charge as a portfolio with assets comparatively more subject to environmental risk, provided that the nature and the direction of the two portfolios are equal.
Secondly, this increase in risk weights would most likely have some unintended consequences, such as an arbitrary increase in the capital charge faced by most banks for the positions in their trading book. Also, it would provide no incentives for them to switch their investment policies in favour of environmentally friendly alternatives.
Thirdly, the calibration of risk weights would be rather difficult to attain without a proper framework and the estimation of such an increase would rely on arbitrary data as environmental risk has not been properly monitored for long periods of time. The calibration of risk weights would need also regulatory inputs and the development of a new set of rules on behalf of the regulator.
RRAO
The discussion paper also mentions RRAO as a possible way of accounting for environmentally related risks. This framework is intended to address risks of complex products or products not fully covered by the SBA and default risk frameworks, but which are already identified by the risk management. Although being identifiable, environmental risk is present in plain-vanilla products and it is difficult to isolate depending on risk factors and asset class. In addition, such an approach will most probably lead to a strong divergence of Market Risk Department and Front Office’s pricing which is against the most recent direction the FRTB is leading us through (i.e., consistency between front office and risk management unit prices).
Concluding remarks for answer 23
To conclude, the approach we consider the most feasible is the elaboration of sub-buckets properly set for environmental risk. Still, we envisage two hurdles which must be overcome in order to set this method into motion. Firstly, sub-bucketing should not lead to an arbitrary increase of the capital charge for banks. Secondly, the regulator should provide financial institutions with proper risk weights, based on reliable data, although AIFIRM realizes how difficult their calibration might be.
Q24: For the Internal Model Approach, do you think that environmental risks could be better captured outside of the model or within it? What would be the challenges of modelling environmental risks directly in the model as compared to modelling it outside of the internal model? Please describe modelling techniques that you think could be used to model ESG risk either within or outside of the model.
As things stand now, the Internal Model Approach developed in accordance to best practices should already incorporate environmental risk, along with the others.However, a major limitation of modelling environmental risk within the IMA is that, as of now, there is no broad-based consensus on how environmental risk might be specifically isolated in either VaR or ES, thereby rendering its direct observation challenging.
As already mentioned in answer to question 23, at the moment AIFIRM does not deem accurate to define a new environmental risk factor. Risk Management may model curves, surfaces, vertices for each asset classes, but it may not unbundle a common risk factor like the environmental factor without seriously affecting the PLA.
Otherwise, considering environmental risk as a non-modellable risk factor seems not to be in line with the original aim of the non-modellable risk factor approach which should serve complex products only. As a result, such approach would cause an unintended penalization of all instruments with environmental intrinsic factors.
Another consideration AFIRM raised is related to the input to the models such as PDs or historical series of prices. They might be adjusted to estimate environmental risk but there are limitations. Firstly, the current regulatory framework does not allow to adjust historical data. Secondly, environmental risk and green assets are only now starting to make their way into the portfolio of investors in a meaningful amount; therefore, historical data is patchy and unreliable making the job of building parameters specifically geared to account for this type of risk extremely challenging.
The discussion paper also suggests incorporating a specific Add-On into the internal model. However, as the discussion paper is scant on details on how this would be implemented, it is difficult to assess it. This difficulty is further compounded by the recurring theme of the unreliability of environmental data and the divergence in how risks other than the ones normally considered are accounted for.
Instead of a change in the IMA regulatory framework, AIFIRM suggests leveraging the stress testing exercise, just like the one recently set up by the ECB.
Q25: Do you have any other proposals on integrating environmental risks within the market risk framework?
We deem the proposals illustrated in the Discussion Paper to be certainly worthy of consideration. Nonetheless, we would welcome substantiated indications from the regulator with respect to the setting and the parameters applicable to each model in this regard to allow for possibly uniform approaches among banking institutions.AIFIRM proposes to consider the inclusion of environmental risk in the Pillar II framework instead of Pillar I to allow banks to capture the institution-specific nature of environmental risks.
In case changes to SA and IMA are applied, AIFIRM proposes to consider a two-way approach:
- keeping a framework as much as possible in line with the one currently in use/design (eventually with a slightly more prudent RWs) and
- designing a framework to provide incentives, by means of capital benefit, for banks willing to embark in a reassessment of their current framework (and having an environmentally safe portfolio).
All in all, one incentive tool that we estimate to be of relevance is the relative lower charge faced by environmentally risk proofed assets (as already mentioned in answering Q23). This point, along with the possibility of maintaining the current prudential framework, modified introducing new environmentally related adjustments, would be an incentive for financial institutions to rely on the new approach (when and if possible).