Response to consultation on Guidelines on limits on exposures to shadow banking
Go back
The guidelines give the impression that shadow banking entities pose a special, specific type of risk. This is not the case. Like any other borrower, a shadow banking entity basically has the potential to generate various types of risk for the lending bank (credit risk, market risk, operational risk, etc.). It is already an established principle in risk assessment that the creditor needs to understand and be aware of the opportunities and risk of the transaction and the activities run by the debtor. Knowing the customer is essential to understand these aspects and this is nothing specific for an SBE.
Also, it is excessive, in our view, to assume that shadow banking entities by their very nature have a correlation of one and thus pose a high level of concentration risk. This automatically puts shadow banking entities on a worse footing than other borrowers, such as corporates, which is not appropriate. It therefore makes little sense to require separate management processes and control mechanisms to be developed for shadow banking entities. Nor would this be feasible, since the shadow banking sector is far too heterogeneous for the application of uniform processes and mechanisms.
Banks are naturally required under Pillar 2 to identify, measure and manage concentrations of credit risk. Applied to shadow banking entities, this means that – as stipulated in Article 81 of CRD IV – the concentration risk arising from a shadow bank’s links to other borrowers (group of connected clients) or through sectoral or geographic concentration has to be adequately managed. However, even the CEBS’s Guidelines of December 2011 on the revised large exposure regime point out that only idiosyncratic risk should be addressed by the large exposure regime itself while geographic and sectoral risk should be dealt with under Pillar 2.
It is normal practice to set sectoral and geographic limits on credit risk under Pillar 2. Individual limits are set as part of banks’ routine lending processes. The “hard” requirements of Pillar 1 and the large exposure regime already have a limiting effect in this context. Given the heterogeneity of the shadow banking sector, we believe it makes little sense to regard shadow banking entities as a single industry and place corresponding limits on a bank’s exposure to them.
In any event, it should be made clear that, in point e), the process to determine the interconnectedness between shadow banking entities should include only the SBE to which the bank has exposures. An exercise to assess the correlation between the bank’s exposures to SBE with other SBE in the market, not included in the bank’s portfolio, would be impossible to perform as no detailed information exists in the bank (neither needs to be collected) for those remaining SBE.
Please also note our general comments attached.
Furthermore, it might be on a too detailed and operational level to put demands on the institutions to “review and approve the management process to shadow banking entities.” It is probably more relevant and sufficient that the Chief Risk Officer and Risk Control function have responsibility of such follow up and just report any deficiencies within the ordinary reporting scheme to the management body of the institution.
Please also note our general comments attached.
Since it seems, however, that policymakers wish to set limits, we would suggest considering the following alternatives:
Blanket aggregate limits for large exposures: this would make it superfluous to set individual limits specifically for shadow banks under Pillar 2. The regular Pillar 2 requirements would continue to apply, including to shadow banks. We believe an appropriate level for the aggregate limit would be 500% of eligible capital. This would mean that a bank’s large exposures to shadow banking entities could not exceed five times the amount of its eligible capital.
Lower individual large exposure limits: in this case we would suggest a general individual limit of 20% of eligible capital.
If the above alternatives are not considered acceptable and the idea of establishing both individual and aggregate limits is retained, we consider it essential to drop the fallback approach.
The determination of interconnectedness between shadow banking entities, and between shadow banking entities and financial institutions, should be conducted only regarding direct linkages: when shadow banking entities form part of the bank credit intermediation chain, are directly owned by banks, benefit directly from bank support or when banks hold assets of shadow banking entities. Indirect linkages, such as for example the investment of banks and shadow banking entities in similar assets, or the exposure to a number of common counterparties, cannot be correctly assessed and therefore should be out of scope.
Improvements in data availability and granularity (capital, leverage, liquidity, portfolio composition, etc.) will be essential for adequately capturing the magnitude and nature of risks in the shadow banking system.
Please also note our general comments attached.
It is understandable that the EBA wishes to create incentives to collect as much, and as complete, information as possible about shadow banking entities. But the proposed fallback approach ignores the materiality aspect which is part of every loan decision. There is no need for a “technical” fallback approach, in our view, because shortcomings in setting limits – whether on exposures to shadow banking entities or any other borrower – can be addressed under the SREP and additional capital requirements will put pressure on banks to eliminate them.
We are not clear on the rationale behind the EBA’s preference for option 1. Naturally, this approach is the most conservative of all possible options. But a limit of 25% on all exposures to shadow banks is without doubt far too low. As things stand, banks may, in principle, lend up to 25% of their eligible capital to each shadow banking entity with which they do business. Option 1 would therefore significantly overstate the risks involved in lending to the shadow banking sector.
Option 1 does not provide incentives to develop a robust assessment process as the non-compliance with a principal approach for just one SBE exposure will lead to an overall limit to all SBE exposures. Even if, in these situations, banks decide to runoff the portfolios to the specific SBEs to which they cannot obtain the required information, this will lead to an increase of exposures to “principal approach compliant” SBE, which, on the other hand, will result in concentration risk challenges. Furthermore, option 1 could lead, on the short term, to swift systemic events resulting from the insolvency/ fire sale of assets from the SBE that cannot provide the necessary information of the banking sector.
The smoothing and mitigation of these risks can only be achieved through the development of specific regulatory regimes to SBE that allow them to redesign their operating/business models, with sufficient time for implementation. Any isolated measure that influence unilaterally banks’ behaviours, without allowing SBE to adapt, will always have the risk to trigger adverse movements, which will be substantially higher with option 1 (the potential deleveraging process will affect not only the specific SBE that do not provide the required information but all SBE entities/markets).
In summary, SBEs are different and will be at different stages of evolution, both features that are not controlled by banks. As such, the potential absence of information from specific SBE groups should not lead to restricted limits to other sophisticated SBE groups, as proposed under option 1.
Please also note our general comments attached.
We do not agree with the assumption mentioned on page 13 which states that “in the absence of sufficient information, all exposures to shadow banking entities could be connected” and so indicates that they should be understood as the same client. As such, although the information gathered could not allow the compliance with all specific rules of the principal approach under Title II, this does not mean that, based on the information collected, there is an inability to evaluate the interconnection between SBE and so, concluding that they are all connected. Moreover, this example also illustrates that the framework does not appropriately capture the risk of the exposure to the SBE thus calling for a different, simpler approach.
Due to the complexity of the principle approach, a comprehensive comparison of the principle and fallback approaches cannot be done making the assessment of the proposed 25 % limit for shadow banking exposures challenging. However, it is of paramount importance that subsidiaries to banking entities under consolidated supervision, as proposed in the draft Guidelines, are excluded from the limits according to the definition of shadow banking entities. Taking into account that the regular large exposure limit of 25 % of eligible capital applies to single clients or single group of connected clients, it very surprising that EBA is of the opinion that an aggregate limit of 25 % would be sufficient. In CRD II there was an aggregate limit for all large exposures (exposures exceeding the 10%-threshold) of 800 % of own funds. From that figure it becomes obvious that the proposed limit of 25% is far too low.
We propose that the 25% limit should be subject to quantitative impact studies that allow the development of an empirical and supported base for the calibration of that factor. As an alternative, we suggest that banks might have the possibility to segment SBE exposures between specific sub-groups for which they can proof that no correlation is observed, even if the remaining requirements are not totally fulfilled to design individual limits for each specific SBE.
Please also note our general comments attached.
2. Do you agree with the approach the EBA has proposed for the purposes of establishing effective processes and control mechanisms? If not, please explain why and present possible alternatives.
We are opposed to the idea of introducing additional Pillar 2 qualitative requirements, like the ones in Title II paras 1 and 2, explicitly for shadow banking entities. The qualitative Pillar 2 requirements set by CRD IV have already been implemented in full in national law. It is not clear why these guidelines should set requirements aimed specifically at shadow banks when the same requirements already apply to all borrowers anyway and are already enshrined in various other EBA guidelines (e.g. those on internal governance, SREP). The requirements in Title II, paras 1 and 2 would create unnecessary additional administrative work since separate frameworks, policies and reporting systems would need to be developed specifically for shadow banking entities and these would be subject to separate scrutiny by supervisors and auditors. There would be no corresponding benefit.The guidelines give the impression that shadow banking entities pose a special, specific type of risk. This is not the case. Like any other borrower, a shadow banking entity basically has the potential to generate various types of risk for the lending bank (credit risk, market risk, operational risk, etc.). It is already an established principle in risk assessment that the creditor needs to understand and be aware of the opportunities and risk of the transaction and the activities run by the debtor. Knowing the customer is essential to understand these aspects and this is nothing specific for an SBE.
Also, it is excessive, in our view, to assume that shadow banking entities by their very nature have a correlation of one and thus pose a high level of concentration risk. This automatically puts shadow banking entities on a worse footing than other borrowers, such as corporates, which is not appropriate. It therefore makes little sense to require separate management processes and control mechanisms to be developed for shadow banking entities. Nor would this be feasible, since the shadow banking sector is far too heterogeneous for the application of uniform processes and mechanisms.
Banks are naturally required under Pillar 2 to identify, measure and manage concentrations of credit risk. Applied to shadow banking entities, this means that – as stipulated in Article 81 of CRD IV – the concentration risk arising from a shadow bank’s links to other borrowers (group of connected clients) or through sectoral or geographic concentration has to be adequately managed. However, even the CEBS’s Guidelines of December 2011 on the revised large exposure regime point out that only idiosyncratic risk should be addressed by the large exposure regime itself while geographic and sectoral risk should be dealt with under Pillar 2.
It is normal practice to set sectoral and geographic limits on credit risk under Pillar 2. Individual limits are set as part of banks’ routine lending processes. The “hard” requirements of Pillar 1 and the large exposure regime already have a limiting effect in this context. Given the heterogeneity of the shadow banking sector, we believe it makes little sense to regard shadow banking entities as a single industry and place corresponding limits on a bank’s exposure to them.
In any event, it should be made clear that, in point e), the process to determine the interconnectedness between shadow banking entities should include only the SBE to which the bank has exposures. An exercise to assess the correlation between the bank’s exposures to SBE with other SBE in the market, not included in the bank’s portfolio, would be impossible to perform as no detailed information exists in the bank (neither needs to be collected) for those remaining SBE.
Please also note our general comments attached.
3. Do you agree with the approach the EBA has proposed for the purposes of establishing appropriate oversight arrangements? If not, please explain why and present possible alternatives.
As outlined in our comments to question two, we do oppose the idea of introducing special qualitative requirements for SBEs only that are usually part of other regulation for pillar II. Risk management processes are reviewed by the management body on a regular basis. But this assessment covers the entire risk portfolio. In addition we would like to draw again the attention on the fact that SBEs should not be considered as a single risk category.Furthermore, it might be on a too detailed and operational level to put demands on the institutions to “review and approve the management process to shadow banking entities.” It is probably more relevant and sufficient that the Chief Risk Officer and Risk Control function have responsibility of such follow up and just report any deficiencies within the ordinary reporting scheme to the management body of the institution.
Please also note our general comments attached.
4.Do you agree with the approaches the EBA has proposed for the purposes of establishing aggregate and individual limits? If not, please explain why and present possible alternatives.
As explained above, we do not believe it makes good sense to set separate limits on exposures to shadow banking entities under Pillar 2 since the shadow banking sector is highly heterogeneous and no risk management benefits would ensue, which should always be a prerequisite for setting a limit. In any event, individual limits are already set for every client and group of connected clients as a result of routine lending processes or banks’ strategies for managing credit risk.Since it seems, however, that policymakers wish to set limits, we would suggest considering the following alternatives:
Blanket aggregate limits for large exposures: this would make it superfluous to set individual limits specifically for shadow banks under Pillar 2. The regular Pillar 2 requirements would continue to apply, including to shadow banks. We believe an appropriate level for the aggregate limit would be 500% of eligible capital. This would mean that a bank’s large exposures to shadow banking entities could not exceed five times the amount of its eligible capital.
Lower individual large exposure limits: in this case we would suggest a general individual limit of 20% of eligible capital.
If the above alternatives are not considered acceptable and the idea of establishing both individual and aggregate limits is retained, we consider it essential to drop the fallback approach.
The determination of interconnectedness between shadow banking entities, and between shadow banking entities and financial institutions, should be conducted only regarding direct linkages: when shadow banking entities form part of the bank credit intermediation chain, are directly owned by banks, benefit directly from bank support or when banks hold assets of shadow banking entities. Indirect linkages, such as for example the investment of banks and shadow banking entities in similar assets, or the exposure to a number of common counterparties, cannot be correctly assessed and therefore should be out of scope.
Improvements in data availability and granularity (capital, leverage, liquidity, portfolio composition, etc.) will be essential for adequately capturing the magnitude and nature of risks in the shadow banking system.
Please also note our general comments attached.
5. Do you agree with the fallback approach the EBA has proposed, including the cases in whichit should apply? If not, please explain why and present possible alternatives. Do you think that Option 2 is preferable to Option 1 for the fallback approach? If so, why? In particular: Do you believe that Option 2 provides more incentives to gather information about exposures than Option 1? Do you believe that Option 2 can be more conservative than Option 1? If so, when? Do you see some practical
If there is really a need for a technical fallback approach, what we doubt, we would support a fallback approach, based on option 2. SBE will be a very heterogeneous group with different business models, levels of disclosure and with different risk levels within their portfolios. Based on this heterogeneity, it does not seem appropriate that, if a credit institution gathers all required information for the majority of those entities but, for a small group of SBE, cannot obtain the information required to set a meaningful limits framework, all the bank’s exposures to all SBE (regardless of the information obtained) should be perceived as an exposure to the “same client” and, as such, will be subject to a 25% aggregate limit.It is understandable that the EBA wishes to create incentives to collect as much, and as complete, information as possible about shadow banking entities. But the proposed fallback approach ignores the materiality aspect which is part of every loan decision. There is no need for a “technical” fallback approach, in our view, because shortcomings in setting limits – whether on exposures to shadow banking entities or any other borrower – can be addressed under the SREP and additional capital requirements will put pressure on banks to eliminate them.
We are not clear on the rationale behind the EBA’s preference for option 1. Naturally, this approach is the most conservative of all possible options. But a limit of 25% on all exposures to shadow banks is without doubt far too low. As things stand, banks may, in principle, lend up to 25% of their eligible capital to each shadow banking entity with which they do business. Option 1 would therefore significantly overstate the risks involved in lending to the shadow banking sector.
Option 1 does not provide incentives to develop a robust assessment process as the non-compliance with a principal approach for just one SBE exposure will lead to an overall limit to all SBE exposures. Even if, in these situations, banks decide to runoff the portfolios to the specific SBEs to which they cannot obtain the required information, this will lead to an increase of exposures to “principal approach compliant” SBE, which, on the other hand, will result in concentration risk challenges. Furthermore, option 1 could lead, on the short term, to swift systemic events resulting from the insolvency/ fire sale of assets from the SBE that cannot provide the necessary information of the banking sector.
The smoothing and mitigation of these risks can only be achieved through the development of specific regulatory regimes to SBE that allow them to redesign their operating/business models, with sufficient time for implementation. Any isolated measure that influence unilaterally banks’ behaviours, without allowing SBE to adapt, will always have the risk to trigger adverse movements, which will be substantially higher with option 1 (the potential deleveraging process will affect not only the specific SBE that do not provide the required information but all SBE entities/markets).
In summary, SBEs are different and will be at different stages of evolution, both features that are not controlled by banks. As such, the potential absence of information from specific SBE groups should not lead to restricted limits to other sophisticated SBE groups, as proposed under option 1.
Please also note our general comments attached.
6. Taking into account, in particular, the fact that the 25% limit is consistent with the currentlimit in the large exposures framework, do you agree it is an adequate limit for the fallback approach? If not, why? What would the impact of such a limit be in the case of Option 1? And in the case of Option 2?
First, the adoption of a principal approach requires not only the gathering of sufficient level of information but also the ability to effective use that information on the definition of a comprehensive set of limits. Second, we highlight that the compliance with the requirements stated in the principal approach include numerous points - from a) to c) for aggregate limits and from a) to h) for individual limits - being just one of them the interconnectedness.We do not agree with the assumption mentioned on page 13 which states that “in the absence of sufficient information, all exposures to shadow banking entities could be connected” and so indicates that they should be understood as the same client. As such, although the information gathered could not allow the compliance with all specific rules of the principal approach under Title II, this does not mean that, based on the information collected, there is an inability to evaluate the interconnection between SBE and so, concluding that they are all connected. Moreover, this example also illustrates that the framework does not appropriately capture the risk of the exposure to the SBE thus calling for a different, simpler approach.
Due to the complexity of the principle approach, a comprehensive comparison of the principle and fallback approaches cannot be done making the assessment of the proposed 25 % limit for shadow banking exposures challenging. However, it is of paramount importance that subsidiaries to banking entities under consolidated supervision, as proposed in the draft Guidelines, are excluded from the limits according to the definition of shadow banking entities. Taking into account that the regular large exposure limit of 25 % of eligible capital applies to single clients or single group of connected clients, it very surprising that EBA is of the opinion that an aggregate limit of 25 % would be sufficient. In CRD II there was an aggregate limit for all large exposures (exposures exceeding the 10%-threshold) of 800 % of own funds. From that figure it becomes obvious that the proposed limit of 25% is far too low.
We propose that the 25% limit should be subject to quantitative impact studies that allow the development of an empirical and supported base for the calibration of that factor. As an alternative, we suggest that banks might have the possibility to segment SBE exposures between specific sub-groups for which they can proof that no correlation is observed, even if the remaining requirements are not totally fulfilled to design individual limits for each specific SBE.
Please also note our general comments attached.