Read InterestingReading_Filename_104.pdf text version

Regulators move to impose AMA capital floors

The advanced measurement approach to operational risk management was devised to ensure firms' op risk capital reflected the true nature of the risks it was carrying. Both the standardised (TSA) and basic indicator (BIA) approaches simply apply an indicator to gross income which is a proxy for the bank's op risk exposure. However, in this new world environment of more intense regulatory scrutiny and the trend towards ensuring banks hold more capital as buffers against further shocks to the financial system, some national regulators are limiting the AMA capital calculation by imposing a capital floor linked to the TSA betas as a percentage. With effect from March 1 this year, the South African Reserve Bank (SARB) imposed operational risk capital AMA floors at a risk category level. South African AMA banks are not allowed to reduce AMA Risk Weighted Assets (RWA) to below 85% of the TSA equivalent. So depending on the business line mix of the bank, this imposes a minimum P1 capital of around 12.75%. Other regulators are also imposing capital floors or are carrying over the transitory caps outlined in the Basel II Accord past the implementation date ­ although they are applying the floor at a total Pillar 1 capital level, not at the business line level, which SARB is now doing. In Australia, the Australian Prudential Regulatory Authority (APRA) imposed a blanket transitional capital floor of 90% RWA based on TSA capital for AMA banks in 2008, which remains in place today. This floor is not based on either business line or risk category differentiation. APRA's Prudential Standard 150, paragraph 5, states: "For a transitional period commencing with implementation of the Basel II Framework (set out in International Convergence of Capital Measurement and Capital Standards: A Revised Framework) in Australia, APRA will limit reductions in minimum regulatory capital allowable to an ADI [authorised deposit-taking institutions] that has IRB and AMA approval, relative to what would have been required had Basel I remained in force. This limit will be implemented by way of a transitional floor on the risk-weighted assets (RWA) figure used for determining the ADI's actual risk-based capital ratio and whether it meets its prudential capital ratio (PCR). The method of calculating the transitional floor adjusted RWA figure is detailed in paragraphs 7 to 12 of this Prudential Standard." Most recently, the Bank of Spain has become the latest national regulator to impose an AMA capital floor for its newly-approved, first and only AMA bank, BBVA. However, this time the floor is set at 100% of TSA. BBVA Group received AMA approval for operational risk in Spain and Mexico from the Bank of Spain in March 2010 however approval was conditional on the bank resolving certain issues before December 2011. Until those issues are resolved, the bank's AMA model is floored at the standardised level. The calculation is performed using AMA but the level of correlation authorized is such that capital does not go below 100% of TSA. BBVA was not expecting this situation but because the floor is transitory the bank is hoping it will start to make capital savings by the end of 2011. Germany and the UK have no such floors in place, although one German head of op risk says that

although the German regulator Bundesanstalt für Finanzdienstleistungsaufsicht (BaFin) does not have capital floors in place, it is "aiming to increase AMA capital wherever it can to grow the money in the system". The French did consider imposing a crude capital floor, as explained by Duc Pham Hi, Professor of Computational Finance; Credit & Op Risk Consultant and ex-regulator, in a post on this subject on the RiskBusiness Operational Risk LinkedIn discussion group page: "Back in 2005-2007 at the French Commission Bancaire, we evaluated a tentative threshold of "acceptability" for AMA as 20%-30% below a crude TSA," he says. As the APRA Prudential Standard 150 indicates, the original notion for imposing capital floors came from Paragraph 49 of Basel II Accord, which advised regulators to implement a transitory capital floor for both the internal ratings-based approach (IRB) for credit risk and also for the advanced measurement approach (AMA) for operational risk. "The Committee believes it is appropriate for supervisors to apply prudential floors to banks that adopt the IRB approach for credit risk and/or the AMA for operational risk following year-end 2008. For banks that do not complete the transition to these approaches in the years specified in paragraph 46, the Committee believes it is appropriate for supervisors to continue to apply prudential floors -- similar to those of paragraph 46 -- to provide time to ensure that individual bank implementations of the advanced approaches are sound. However, the Committee recognises that floors based on the 1988 Accord will become increasingly impractical to implement over time and therefore believes that supervisors should have the flexibility to develop appropriate bank-bybank floors that are consistent with the principles outlined in this paragraph, subject to full disclosure of the nature of the floors adopted. Such floors may be based on the approach the bank was using before adoption of the IRB approach and/or AMA." Paragraph 46 states: "The capital floor is based on application of the 1988 Accord. It is derived by applying an adjustment factor to the following amount: (i) 8% of the risk-weighted assets, (ii) plus Tier 1 and Tier 2 deductions, and (iii) less the amount of general provisions that may be recognised in Tier 2. The adjustment factor for banks using the foundation IRB approach for the year beginning year-end 2006 is 95%. The adjustment factor for banks using (i) either the foundation and/or advanced IRB approaches, and/or (ii) the AMA for the year beginning year-end 2007 is 90%, and for the year beginning year-end 2008 is 80%." Some national authorities have opted to keep the transitory capital floors from the original Basel II document past the implementation date. The Monetary Authority of Singapore also took this option. In December 2010, MAS decided to make the 80% transitional capital floor, which was initially meant to be in place for three years, permanent. Although this has had a limited effect on operational risk as no banks in Singapore are AMA. The section highlighted below in MAS 637 was added in December 2010: "5.1.2A For the third year and all subsequent years from the date of adoption of the IRBA, a Reporting Bank which has adopted the IRBA shall meet capital floors at both the Solo and Group levels of the Tier 1 Capital Resources Requirement and Total Capital Resources Requirement calculated using the approaches adopted by the Reporting Bank under this Notice, and ­ (a) 80% or such other percentage (not

more than 100%) as the Authority may require, of the Tier 1 Capital Resources Requirement and Total Capital Resources Requirement respectively, calculated using the rules in MAS Notice 637 in force immediately before 1 Jan 2008; or (b) 80% or such other percentage (not more than 100%) as the Authority may require, of the Tier 1 Capital Resources Requirement and Total Capital Resources Requirement respectively, calculated using the rules in this Notice with SA(CR) for its credit risk capital requirement5, whichever is higher. 5.1.3 The Authority may impose additional bank-specific capital requirements, including extending the period of operation of the capital floors set out in paragraph 5.1.1, 5.1.2 and 5.1.2A." One regulatory source says most regulators during the initial implementation period of Basel II did entertain some form of capital floor. "Most regulators were applying capital floors with respect to the potential capital reductions obtained through AMAs," he says. "Some of them applied these capital floors based on BIA but the majority used TSA/ASA as the basis. The foundation for applying these capital floors and keeping them after the transitional period was not only paragraph 49 in Basel II but also national supervisory discretion resulting from onsite AMA reviews, validation efforts, quality of ICAAPs, SREPs and the behaviour of AMA models during the financial crisis." One op risk expert in Asia says: "The imposition of transitional capital floors is common as the robustness of the internal models in generating consistent and predictive estimates needs to be proven over time. Additionally, at the point the regulators approve the adoption of the models there could be a number of areas for improvement (e.g. improvement in systems and processes) which the regulators want the bank to address and as such the regulator could impose a higher bank-specific capital floor on a bank. The higher floor would then be lifted if the bank met the conditions imposed by the regulator." This is the situation facing BBVA at the moment. Developing and implementing a robust AMA model is not easy and regulators have a duty to ensure that those banks using the AMA model to calculate their own op risk exposure and level of regulatory capital are getting it right. Putting in place a capital floor linked to the TSA betas is in effect putting the bank into a parallel run state, which can be lifted once the model has proven to be effective. However, imposing such a floor for an indeterminate length of time, like in the case of South Africa, has caused some consternation. One manager says: "We don't believe it is within the letter or spirit of Basel II and watch these developments closely. [Applying] an 85% flooring of AMA op risk ... intuitively feels wrong." Post-financial crisis, regulators have indicated that they think op risk capital levels are too low, especially since data from the mini-loss data collection exercise conducted as part of Basel Committee on Banking Supervision's (BCBS) QIS 6 exercise showed levels of op risk losses were increasing even though capital numbers stayed the same. Imposing an AMA floor is one way to raise capital levels, another way is via Pillar 2 in the Internal Capital Adequacy Assessment Process (ICAAP). In some cases several European regulators have required banks to hold additional top-up capital over and above both TSA and AMA levels through Pillar 2.

The capital floor is usually set under Pillar 1 but it can vary depending on national discretions. One European regulator says: "If the TSA capital charge is felt to be too low and the AMA capital is found to be linked to (eg lower than) the TSA, a 100% floor might be appropriate. We all know that BIA and TSA charges are based on an indicator which is only a proxy of the op risk exposure; as such it may be not able to adequately reflect the current bank's op risk experience, while the AMA should do. I would not be surprised if regulators found (indeed they have found in some cases) that the AMA is higher than the TSA charge. A top manager should ask the op risk manager which is the appropriate AMA capital to cover the bank's op risk exposure rather than asking which is the saving of capital stemming from the adoption of the AMA." One op risk manager in the Asia-Pacific region debated the effect of imposing a capital floor at the Pillar I or the Pillar 2 level: "It is an interesting point about whether the floor is imposed under Pillar 1 or Pillar 2. Under Pillar 1 a multiplier is applied at the discretion of the regulators, while Pillar 2 is an assessment of the quality of the firm's op risk management. Some regulators apply a points system for Pillar 2 increases while others apply a multiplier. For the AMA, I suspect the capital floor imposed under Pillar 1 would be lower than under a Pillar 2 multiplier." He adds: "As long as the regulators peg a multiplier/floor to op risk, capital levels will go up and down with the tide." He gives one extreme, hypothetical, example of an AMA bank with a very good [read low] op risk capital model but, he warns, if the bank was found to have substandard risk management processes in place it would get "hammered" under Pillar 2. "You will find banks are cagey on what their actual op risk capital numbers are precisely because of the fact they don't want their competitors to find out how they were treated by the regulators under Pillar 2," he says. One global head of op risk at an AMA banks says imposing a capital floor at the Pillar 1 level is a problem because it is a public figure. "If a capital floor is imposed at the Pillar 1 level as a regulatory capital buffer it could have profound implications for the bank as the capital ratios are published. However, if it is applied in Pillar 2 it isn't really any different to now, where your capital calculations are added to by the regulator to a level where they are comfortable with the capital you are holding." Critics of capital floors suggest it would be better to fix all that is wrong with the op risk section of the Basel II Accord rather than impose floors that are linked to TSA or BIA to act as a sticking plaster until the BCBS gets it act together to make some serious changes. That said, it could also be a sign of things to come. The global head of op risk commented that the news that South Africa and Spain have imposed capital floors relating to the TSA is not necessarily "the direction of travel, but there are some concerns among regulators that the op risk capital is too low ­ not just AMA but op risk in general". He adds: "[Regulators] are questioning particularly whether the TSA betas are set too low. SIGOR is reviewing the whole debate and asking whether, post-crisis, they are set at an appropriate level.

Both measures for credit and market risk have risen, but op risk has not. It is unsurprising therefore that they are looking at op risk and if the regulatory community is agreed that op risk capital levels are too low, the fact that some regulators are imposing some kind of floor is not completely surprising and it is within the national discretion allowance in Basel II." The Standards Implementation sub-Group on Operational Risk at the BCBS (SIGOR) has a working group in place that is debating whether the TSA betas should be raised and by how much and using which indicator. These discussions have been rumbling on for more than two years now but the working group was only set up last year to look seriously at how the betas need to be changed. The progress is slow, however and given how busy the individual regulators are away from SIGOR, any changes are unlikely to emerge until 2012. One SIGOR member says: "It is unavoidable that the SIGOR is currently investigating the BIA/TSA op risk discipline to assess if enhancements are necessary or not. However, the process is not expected to be quick. It will require time in order to have, in the end, an outcome of higher quality. No directions on what the SIGOR outcome should look like may be disclosed at this stage." The last loss data collection exercise showed that AMA banks held a much lower percentage of op risk regulatory capital than BIA or TSA, which is what many would expect. However there was the feeling among regulators that it was too low. If the betas are raised by SIGOR, and the AMA has a capital floor in place, the AMA will rise too. But the question this throws up is, if a capital floor of 100% of TSA is imposed on AMA, why would banks bother to spent time and resources on regulatory capital modelling? They may want to model their operational risk for economic capital purposes but they probably don't need to model to the same standard that is required in the AMA for regulatory capital. The global head of op risk says: "In a way I can see the regulators' desire to impose a capital floor but they need to be careful where that floor is set. Too high and they risk discouraging development in operational risk measurement techniques. Op risk measurement and modelling is behind credit and market risk already, this would set it back even further." He adds: "I would say that an 85% floor would still provide enough incentive to go AMA and encourage development of op risk measurement and good practice. This would give some lassitude to develop a model. I didn't say it would be an acceptable level ­ it is only the level where the AMA would remain worth the effort. At 85% all you are doing is parallel running the TSA at that point." Sören Eng, Senior Credit Risk Analyst at Skandinaviska Enskilda Banken, commented on the LinkedIn RiskBusiness discussion group page that the move towards imposed AMA capital floors will just be another reason for banks not to go for the advanced approach: "How many AMA banks are there? I think this will create one more excuse for banks to remain standardised, as the incentive will disappear. SEB was one of the first financial groups to go AMA. It has granted us a good capital relief, but no improved rating. With respect to rating, BFSR - basic financial strength is all that matters, thus, this will just be a blow in the air!"

Not all op risk managers are against the idea of a capital floor, however: "I guess some visibility over floors would actually be helpful," says one UK-based op risk manager. "I think we have always assumed that there would be a limit below which a bank wouldn't be allowed to reduce capital under the AMA compared to TSA but this level has never been explicitly stated. Banks have spent a lot of money on implementing the AMA to operational risk management for what they perceived to be the dual benefit of improving their assessment of the risks inherent in their organisations while also, in most cases, helping to reduce their regulatory op risk capital. Take that capital benefit away, the advantage of the AMA remains only in having a better understanding of operational risk, which helps improve the management of it. One head of op risk at a European AMA bank says: "Although capital savings are a big driver of AMA models, especially in current conditions, I think that the main driver must be operational risk management, operational loss reductions and preventing and anticipating significant events that could endanger a bank's solvency or continuity. However, it is true that the AMA entails a compromise that costs money in terms of systems and human resources, and capital savings are easy to quantify as a benefit of AMA, whereas management op risk benefits are very difficult to quantify." Better and more effective operational risk management is ultimately what most op risk managers aspire to as the source above notes it is extremely difficult to convince chief executives for more resources to manage operational risk without the cuts in capital to back it up. Mike Finlay, chief executive of RiskBusiness International, says: "The reality is that an AMA model does not improve risk management above the TSA levels and that a model itself is both unrealistic and difficult to comprehend. Enforcing capital floors linked to TSA remove much of the incentive for firms to move to AMA, so if you can't get reduced capital or insurance relief of capital, why bother?" Another factor that is not yet known is whether a bank that has an AMA model with discounts for expected loss and insurance can apply those discounts to below the capital floor. The global head of op risk source says: "If not, then why bother expending time and effort and the on-going cost to maintain these calculations?" Another manager says it might still be worth it: "The calibration of a model is combining plenty elements without predefined weights ­ I guess if your calculation before expected loss or insurance is at or above TSA it may still make sense." One source in Singapore says that even though no banks are AMA at the moment, any interest in going AMA in the future is waning: "The interest in going AMA initially was because banks thought that they could get a reduction in op risk capital but in the wake of the financial crisis and the new Basel III rules, banks realise that regulators now want banks to be better capitalised (and with higher quality capital too). Interest in AMA is waning because the implementation costs are very

high and banks are not sure if the benefits would outweigh the costs. I think regulators also know this and so they are moving to impose the AMA qualitative criteria on TSA banks [like the UK FSA has with its paper, Enhancing frameworks in the standardised approach to operational risk]. Perhaps they see that it is more important to focus on managing op risk than to be able to measure it more precisely. Managing op risk is more proactive. When a bank needs to rely on its op risk capital to absorb the shock from a huge op risk event it could be already too late. The bank would also have suffered huge reputational damage. Hence, between investing in better systems and processes for managing risks and investing in systems and processes to measure the risk more precisely for capital purposes; the former makes more sense." But they add: "Having said that, I think banks should use AMA approaches to measure their op risk for economic capital purposes. The BIA and TSA are not risk sensitive and do not incentivise businesses to better manage their op risk. We need a risk sensitive method to allocate op risk capital to the business lines. But I agree that, without regulatory push, the development and continued evolution of the discipline of op risk measurement will decline." Although there are problems with the AMA model, there is another benefit for having a model that has been approved by your regulator, even if there is no longer any significant capital benefit. One UK-based op risk manager says: "The threat of Pillar 2 add-ons is likely to see some larger TSA banks move to AMA as a more of a defensive measure as they can better argue against a Pillar 2 add-on by using a model approved by the regulator. This is probably a subtle change from when the AMA was first available to banks that saw it as an opportunity to reduce capital." For the same reason, one European op risk manager of an AMA bank is using a formula presented by the Japanese Financial Services Authority to calculate an operational risk capital benchmark, which he is able to use as a defence against regulators imposing any Pillar 2 add-ons. 1 The Asia-Pacific head of op risk says that although having an AMA model is not essential, advanced management of op risk is, specifically to protect the bank and fend off Pillar 2 capital loading: "Anyone can get an AMA model from a third party provider if they have the data­ you don't need to be an AMA bank to measure your op risk capital number. The challenge is around organisational change to become proficient under Pillar 2." He adds: "If an organisation seeks to pay low op risk capital to save costs then they will be negatively impacted on operational risk capital anyway ­ it's a circular argument. Cut costs on op risk management and pay later. The losses from op risk events are still much higher than total of credit risk and market risk losses, as the crisis has shown. Banks are still in denial about this as credit risk is cheaper to account for in regulatory capital. There is an on-going discussion about the equilibrium is right between credit risk, market risk and operational risk. RWA calculations will fundamentally change. Basel II spent five years gearing up for a clear capital regime to change it now is a major, but an appropriate, thing to do. We have a better insight into operational risk now.


"A simple formula for operational risk capital: A proposal based on the similarity of loss severity distributions observed among 18 Japanese banks", by Tsuyoshi Nagafuji, Takayuki Nakata and Yugo Kanzaki. (May 2011).

Since we were AMA accredited we made a lot of progress in expanding op risk methodologies and projects. Five years on, we are all much more aware ­ we should now have the base infrastructure in place to support an enterprise-wide risk management approach. The hard part is changing the organisation to understand the risks it is taking and assessing how much capital to hold, while still making a return. Linking risk to strategy is fundamental to the future development of the discipline." Moving more large institutions towards the more advanced approach for operational risk management remains a key part in pushing forward to that goal of embedding risk management into the business strategy, which since the financial crisis, is also the aim of most national regulators. However, as demonstrated by the attitude of op risk managers commenting in this paper, imposing capital floors may not be the best way to encourage that aim.

The content of this document is the property of RiskBusiness International Limited. It is made available on the understanding that no part of it shall be modified, copied, stored in a retrieval system, or transmitted in any form, by any means or supplied to a third party without prior written consent of RiskBusiness. Care and attention has been taken in the preparation of this document but RiskBusiness shall not accept any responsibility for any errors or omissions herein. Any advice given or statements or recommendations made shall not in any circumstances constitute or be deemed to constitute a warranty by RiskBusiness as to the accuracy of such advice, statements or recommendations. The entire content has been derived from publicly reported sources and RiskBusiness cannot vouch for the authenticity or correctness of such reports. Any opinion implied, suggested or provided is the considered view of the publisher only. RiskBusiness shall not be liable for any loss, expense, damage or claim arising out of the advice given or not given or statements made or omitted to be made in connection with this document. RiskBusiness recognises copyright, trade marks, registrations and intellectual property rights of certain third parties whose work is included or may be referred to in this document. The content of this document does not constitute a contractual agreement with RiskBusiness. RiskBusiness accepts no obligations associated with this document except as expressly agreed in writing. The information contained in this document is subject to change. All rights reserved. RiskBusiness International Limited, 2 Claremont Way, Halesowen, West Midlands, B63 4UR, United Kingdom. 2011 RiskBusiness International Limited


8 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate


You might also be interested in

Microsoft Word - Nomura - Basel II Impact_09202005.doc