BCBS consultative document: Capital floors: The design of a framework based on standardised approaches
Joint comments from Finanstilsynet (Financial Supervisory Authority of Norway) and Norges Bank. Letter of 25 March 2015 to Basel Committee on Banking Supervision.
Finanstilsynet and Norges Bank appreciate the opportunity to provide our feedback on the proposal for a new capital floor. We share the Basel Committee's concerns about low and diverging capital requirements among banks using internal models to calculate their capital requirements, and recognise the usefulness of floors in the capital requirements as a means to constrain the effects of models. Furthermore, we recognise the need to revise the floor rules as the current floor, based on Basel I, is not sufficiently risk sensitive, and imposes additional costs on banks and regulators having to calculate capital requirements based on out-dated regulation. However, with a more risk-sensitive standardised approach as the basis, the proposed new floor will inherit some procyclicality.
In order to function as a credible backstop to the models, it is important that the floor is reflected in the solvency ratios and buffer requirements. Hence the floor has to apply to the risk weights or to risk weighted assets (RWA), as in the 2006 standards, and not only to the own funds requirements. If the floor, as observed in some European countries, is only applied to the total own funds requirement, the increased capital and buffer requirements in Basel III may be offset by RWA reductions.
In Norway, the Basel I floor increases the capital and buffer requirements for banks using internal ratings-based (IRB) models for credit risk by around 15%-20% on average, and more than 40% for some banks. Hence the floor has been an important factor for the banking sector's capital increase of NOK 116 billion (55%) since 2010.
Floors should not, however, be the preferred or only measure applied to mitigate unintended consequences of model use. Thorough supervisory assessments of the banks' assumptions, data and methods, and close monitoring in order to ensure that the parameters and resulting capital requirements reflect the risk in the portfolios, are essential. Furthermore, we acknowledge that not all exposures and risk parameters are suitable for modelling. Therefore, we welcome the Basel Committee's other initiatives to identify and address low and diverging IRB risk weights, including ongoing monitoring exercises and measures to narrow the modelling options as mentioned in the report to the G20 meeting in November 2014.
Level of aggregation (Q1)
Given that the floor will be calibrated according to aggregation level, Finanstilsynet and Norges Bank believe that floors for each major risk category, i.e. credit risk, operational risk and market risk, are most appropriate. Our main concern is whether IRB models result in prudent capital requirements for credit risk, and we acknowledge that similar concerns would arise regarding market and operational risk models if our banks were more exposed to those forms of risk and models were more widely used in those areas. Floors on the exposure class level would address special concerns about specific types of exposures more accurately, but this approach would also be more complex. In general, we think concerns on the exposure type level are best addressed by constraints on models or parameters, but would support a floor on the exposure type level if this would be more appropriate. Albeit simple and transparent, a floor on the aggregate capital requirement/RWA may be diluted where banks have high capital requirements for some risk categories, for example due to particular risks in the bank or conservative modelling. To summarise, we think floors on capital requirements and RWAs for each major risk category is most appropriate to address our main concerns and also provides the best trade-off between simplicity and precision.
Adjustments for differences in the treatment of provisions (Q2)
Finanstilsynet and Norges Bank agree that adjustments are needed to account for the different treatment of provisions and that these adjustments need to be revised if the reference for the floor is changed. We have no strict preferences among the options presented, but option 2, adjusting the floor by "RWA equivalents" of the capital adjustments, is similar to the method we use today. Option 1 may seem more straightforward in terms of how the treatment of provisions affects the capital base, but this method implies that both the capital and the RWAs, i.e. both the numerator and the denominator of the solvency ratio, have to be adjusted for the floor.
Finanstilsynet and Norges Bank support a revised floor as one of several measures to address low and diverging capital requirements. To function as a credible backstop to internal models, the floor must be reflected in the banks' solvency ratios and buffer requirements as well as the minimum capital requirement. This means that it must apply to the RWA. Furthermore, we must emphasise that the floor is a backstop and cannot replace strict requirements to both the banks' modelling and the supervisory assessments.
Erik Lind Iversen
Ida Wolden Bache