Consultation on draft RTS concerning the assessment of appropriateness of risk weights and minimum LGD values

EBA/CP/2021/18

1 Summary of comments

• There must be rules how risk is determined, how the change in risk is quantified and how this affects risk weights and LGD parameters. The change in risk must be calculated by the authorities in a transparent way..
• Multiple overlapping capital requirements must be avoided, and this must be clearly taken into account in the RTS.
• Decisions must be predictable and there must be a concrete possibility for the bank to exercise its right to effective judicial protection, i.e., to challenge the decision and its reasoning in a court of law, if necessary.
• Climate related risks should not be taken into account in this RTS since there are other legal initiatives dealing with this subject. At this point there is not enough reliable information available. Information should be gathered and analysed first.

Finnish banking industry represented by Finance Finland (FFI) welcomes the RTS. It is very important that the possibility given to the authorities to change the values of risk weights directly (in case of a bank using the standard approach) or indirectly through LGD values (in case of an IRBA bank) must be meticulously regulated.

The subsequent Delegated Act gives the authorities power to change the substance of the law (Capital Requirements Regulation) agreed by the legislators. The use of this power cannot be fully discretionary, and it thus must be restrained appropriately. The decisions must be based on transparent and carefully drafted calculations.

Requirements for authorities should be such that the decisions are predictable, that is, banks should be able to anticipate the possible changes of risk weights or LGD parameters themselves. Only this way banks can take the changes into account in their business planning at an early stage.

There are already several regulations and capital requirements which potentially overlap with the scope of this RTS. Among these are capital requirements for systemic risk, O-SII/G-SII buffers and countercyclical buffer. There may be also national macroeconomic measures in place. There cannot be multiple requirements posed to the same risk and thus it is essential that this is made clear in the RTS.

There should be proper reasons for changes of risk weights or LGD parameters and those should be justified in a way, which makes the exercise of the right to effective judicial protection possible, if necessary. There should be adequate requirements to which the authorities’ decisions are based so that reasons for changes in risk weights and LGDs can be presented clearly and in sufficient detail. This means that decisions are e.g. based on objective criteria that show explicitly that there are no multiple capital requirements posed to the same risk.

Only those pieces of information that can be estimated and measured based on reliable and available data should be considered when decisions are taken about risk weights and LGD parameters. This means that, at least at this point, climate related risks should be left outside of scope of this RTS since sufficient and reliable data is not available and cannot be gathered.

2 General comments

Systemic risk and financial stability are already taken into account by multiple tools set by the authorities. There should be a strict rule in the RTS that if such tools are already used to tackle the identified systemic risk, it is not appropriate to set additional tools to curb the same risks. In order to achieve this there must be a requirement that authorities’ decisions are based on a recognition of risk and risk calculations. Authorities must open up the calculations that lead to the change in risk weights or in LGD values. Otherwise, the RTS and subsequent delegated regulation is in danger of being noncompliant with EU law, which requires that the bank must be in a position to ascertain the reasons upon which the decision is based and must be able to recalculate the RWA amount. This legal principle, which in our opinion also applies here, was made clear in the decision in Case T 411/17 of the General Court of the European Union:

“(87) It is settled case-law that if the judicial review guaranteed by Article 47 of the Charter is to be effective, the person concerned must be able to ascertain the reasons upon which the decision taken in relation to him or her is based, as much as to enable that person to defend his or her rights under the best possible conditions and decide in full knowledge of the circumstances whether it is worthwhile to bring an action before the court having jurisdiction as to put that court fully in a position to exercise its power under the FEU Treaty to review the legality of that decision—”.

Without granular and clear knowledge about risk on which the decisions are based it is also impossible to know whether the same risk has already been taken into account as required in recital 26 of Directive 2019/878:

“–In particular, the relevant competent or designated authorities should duly consider whether measures taken under Article 133 of Directive 2013/36/EU duplicate or are inconsistent with other existing or up-coming measures under Article 124, 164 or 458 of Regulation (EU) No 575/2013.”

For instance, in Finland, many macroprudential tools are already in use. For housing loans, FIN-FSA has set a loan to collateral cap (LTC cap). In addition, a minimum 15 % floor was previously set to IRBA banks’ housing loan portfolios. Furthermore, the Finance Ministry is now planning to introduce debt-to-income limits and many more macroprudential tools, in order to curb household’s indebtedness. Systemic risk buffer and O-SII buffers have been actively used in Finland, as well. That has had a substantial effect on banks capital requirements.

With all these tools already in place, any new ones should be carefully considered and duly and comprehensively justified to the institutions. The RTS rightly states that the authority must assess whether other macroprudential measures in force already address the identified systemic risks. However, this leaves room for consideration and the wording should be more explicit to limit the discretion.

Since the eventual delegated regulation will give more accurate substance to the articles 124 and 164 of the regulation 575/2013 (CRR), it is clear that the RTS cannot leave the rules of the regulation vague and indeterminate. This kind of discretion cannot be left to the authorities without at the same time regulating how effective judicial protection has been secured.

In the RTS the conditions which the relevant authority must consider are quite vague: it is enough that the identified inadequacy of the current risk weights and/or minimum LGD values could adversely affect the current or future financial stability of the Member State. The future financial stability is very difficult to measure or quantify, and supervisory tools should not be based on such a soft assumption. Indeed, the financial stability issues should be tied to loss experience and capital requirements should change when underlying economic conditions actually change, not beforehand based on some vague assumptions that may or may not realize. Because of these reasons, authorities’ decisions will be largely unpredictable if necessary, requirements suggested in our response are not taken into account.

We also point out the clear lack of consistency in legal framework. Requirements posed to banks to internally assess their risks (ICAAP) are very extensive and these are supervised by the authorities. However, there seems to be almost non-existent corresponding requirements to authorities. The courts are supposed to be the instance to control the exercise of the power of public authority. The control is, however, not possible, if authorities can de facto alter the substance of the law itself (i.e. change the RWs or LGDs determined in law) without proper restrictions or prerequisites about the circumstances and reasons for making the changes. The substance of a law cannot be changed by free discretion of the authorities.

The limits defined in the CRR for risk weights or LGDs de facto mean that there is no factual limit of how high requirements can an authority pose to a bank. They could be so high that the bank may breach its regulatory capital requirements and thus fail. Without clear and sufficiently unambiguous rules there is no way to challenge authority´s decisions in a court of law. Hence there would be no guarantee that this kind of power is not misused.

The change in risk weights or LGDs should be based on actual need and there should be evidence that losses in the case of a debtor’s default will rise compared to bank’s loss experience because of the change of circumstances mentioned in the RTS. The bank should have sufficient knowledge of those facts and evidence to which the decision is based to make it possible to exercise the right to judicial protection and effectively challenge those facts and evidence in court, if necessary.

Technically we suggest that there are at least a certain kind of “buckets” or steps how much risk weights or LGDs can change if the underlying facts are changed. There should thus be at least some rules to give clear indication beforehand to a bank whether authority’s decision can be expected and what that decision might be.

3 Comments to the questions in the consultation paper

3.1 What is the respondents’ view on the types of factors to be considered during the determination of the loss expectation for the appropriateness assessment of risk weights under the SA?

The uncertainty about loss expectation of immovable properties comes especially through the risk associated with the collateral. All factors related to the characteristics of the immovable property markets and the riskiness of the exposures need to be considered for exposures secured by mortgages on residential property or commercial immovable property. In addition, the location of the collateral and national laws, for example loans to housing companies which are addressed to the owners of the flats, affect the riskiness and loss expectation.
Macroeconomic key variables are troublesome in adjusting the loss expectations upwards or downwards because it would probably require an on-going evaluation by the regulator whether the reasoning for minimum LGD values still holds.

3.2 What is the respondents’ view of the option of considering climate related risks in the determination of the loss expectation where the relevant authority was in a position to perform an appropriateness assessment to one or more parts of the territory of the Member State? What would for the respondent be the benefits and the challenges (costs) of such option?

It is unclear to which extent climate related risk would affect institution specific risk weights. Climate risks may affect the institution’s portfolios in complex ways, or they even may not have any effect to a particular portfolio at all.

It is too early to consider climate related risks in the determination of loss expectation: data on these types of risks is still limited and international/national climate related policies are still being negotiated.

We are thus in an opinion that data and information should be gathered more, and comprehensive impact study made before there may be any conclusions how climate related risks should be taken into account and what is the right regulative response to it.

3.3 What is the respondents’ view on the conditions when assessing the appropriateness of minimum LGD values?

Assessing the market conditions etc. is relatively straightforward, but the open question is how the assessed systemic risk would be introduced as part of a specific LGD model. The LGD estimates should be based on institution’s loss history, and thus affected by the lending standards, conditions, collateralization, and recovery policies the institution is using. Of course, recovery policies are also affected by the national law. It is not clear how the supervisor would introduce these different aspects to institution-specific LGD model with a simple minimum LGD estimate. If these are introduced, it should be clarified how this is done in a way which will allow transparent comparison between different EU countries.

3.4 What is the respondents’ view on the considerations to be taken into account when assessing the appropriateness of minimum LGD values?

Benchmarking to other member states, although interesting, is quite troublesome. Member states have very different lending standards, and the real estate market can also differ considerably. This may make it necessary to consider the market specificities which then also makes it an absolute necessity to include in the RTS that authorities must open the reasoning for the decision including what is exactly the risk according to which the risk weights or LGD parameters must be changed and the quantification of the risk and how it is transformed to risk weights and LGD parameters. Without this information the legal means to exercise the rights to effective judicial protection cannot be achieved.

3.5 What is the respondents’ view on the use of other data sources?

Additional reporting should be avoided. Banking sector is already producing massive amounts of data to supervisory authorities and to the central banks. That existing data should be utilized instead of introducing new reporting requirements.

3.6 Do respondents want to raise other considerations relevant for the application of this article?

In IRBA models, MoC (Margin of Conservatism) is always considered. If uncertainty in LGD estimates, for example regarding climate risks, is already considered in the MoC component, it should not affect the minimum LGD level. Alternatively, it could be considered, that some systemic risks are reflected in MoC, which would also be more precise for example regarding areal distribution of loans.

FINANCE FINLAND
Veli-Matti Mattila
Director, Chief Economist