Regulatory reform for Markets – BCBS and APRA
Sean Carmody, Executive General Manager, Risk and Data Analytics - 2018 ISDA Annual Australia Conference, Sydney
I would like to thank Scott and Keith for the invitation to speak this afternoon at today’s ISDA conference. My name is Sean Carmody and I am responsible for APRA’s Risk and Data Analytics division. This division encompasses all of APRA’s specialist risk teams, including market risk, as well as the teams responsible for APRA’s data collection and analysis. The teams cover all of the industries APRA supervises: banks (and other deposit-taking institutions), insurance (including general, life and private health) as well as superannuation.
Today I will discuss regulatory reforms affecting financial markets businesses, focusing specifically on APRA’s domain: prudential regulation. But before getting onto the regulation of financial markets, I thought I would pause to consider the aims of prudential regulation.
Regulation of financial services has been under intense focus in Australia this year and the varying mandates of different regulators have not always been clearly understood. The primary role of a prudential regulator is to ensure the safety and soundness of financial institutions. For APRA, this means that we aim to ensure that the entities we supervise are able to make good on their financial promises so banks will be able to provide their depositors with access to their funds, insurers will have the financial resources to pay valid insurance claims and superannuation funds will manage money in way that is consistent with the objectives of their members. Not surprisingly, this means that much of the attention of prudential supervisors is focused on financial indicators of safety, such as capital and liquidity. However, we also consider broader sources of risk. Over recent years, APRA has increased its supervision of remuneration practices, governance and even culture. In each instance, our aim has been to ensure that institutions are managing their risks in a way that supports their long-term financial soundness. Earlier this year we published the CBA Prudential Inquiry Report. It provides a powerful case study illustrating the relevance of conduct, remuneration, governance and culture issues to enterprise risk management. So, while APRA is not a conduct regulator, a market integrity regulatory or a financial crime regulator, we will always have an interest in the potential for issues in these domains to have prudential implications.
From that broad perspective, coming back to financial markets, today I will be focused on the traditional concerns of a prudential regulatory. Even with this narrow focus, APRA’s regulatory framework for derivatives and markets spans many dimensions ranging across liquidity risk, funding, leverage, large exposures, valuation and the various “XVA” valuation adjustments, margining, risk mitigation and regulatory capital requirements for default risk, credit valuation adjustment (CVA) risk and, of course, traditional market risk. The regulatory landscape has already changed significantly in the wake of the financial crisis 10 years ago and there is more change still in train. The general trend of these changes in prudential regulation has been towards greater risk sensitivity in standardised approaches, increased scepticism of internal models and greater complexity in implementation.
Since I don’t plan to talk for the rest of the afternoon, I will be selective in the topics I cover and focus on some of the more recent policy initiatives and look ahead at some upcoming changes both at the APRA and Basel level.
One of those recent policy initiatives is margining for non-centrally cleared derivatives. This has been a complex task, involving changes to Australian legislation, a significant re-documentation effort, the development of a single industry model for initial margin and the development of many new systems, processes and pipes. Despite the scale of the task, implementation across the industry has so far gone relatively well. We have had a few major Australian banks phase in for initial margin (IM) under prudential standard CPS 226 and the model approval processes for the Standard Initial Margin Model (SIMM) have gone fairly smoothly. This is not to say that the implementation process hasn’t been without its issues, but it is an encouraging confirmation that complex and significant regulatory change can be implemented well if given sufficient time and managed effectively by both industry participants and regulators.
One of the issues that did arise was the complexity of cross-border application of margining rules. The provisions for substituted compliance such as those in CPS 226 have not been as useful or as widely used as we had originally intended. The use of the stricter rule now has widespread use across the industry but we hope that, as initial margin becomes better established, banks will re-examine the use of substituted compliance particularly where there are differences in margining rules. Another issue on our forward agenda is the significant implementation burden expected for Phase 5 initial requirements, which is scheduled for September 2020, and the resulting challenges many smaller firms will face in implementing the initial margin requirements. This is an issue which we have discussed with ISDA and I is one which we and other regulators around the world will continue to grapple with and discuss with industry. Our own approach will continue to be balancing alignment with international standards to ensure appropriate management of prudential risks as they arise in Australia with supporting competition. This being sensitive to the regulatory burden on industry, particularly for smaller entities.
APRA has also recently finalised the standardised approach to counterparty credit risk, SA-CCR, and the adjusted current exposure measure (CEM) which constitutes a significant refresh to the simple counterparty credit risk exposure framework that we had in place for a very long time. The broad aim we had was to move the larger banks, which have much more significant derivative activities, onto SA-CCR and keep the changes for the smaller banks to a relatively small tweak. As part of the package, we also updated the treatment of bank exposures to central counterparties (clearing houses) and enhanced major bank reporting requirements on their counterparty credit risk exposures. The resulting richer dataset will allow us to develop a much deeper understanding of major banks’ counterparty credit risk exposures and we hope to see significant improvements in our data analytics capability in this area that will assist us in our prudential supervision.
With a July 2019 implementation date, we recognise that we are ahead of many other major jurisdictions on SA-CCR, but perhaps we are a little different in that we do not allow the Internal Model Method (IMM) in Australia. And while SA-CCR is used in other parts of the framework, this earlier adoption has not posed any significant concerns for our larger banks. Derivatives form a relatively small proportion of exposures from a leverage ratio perspective and since we do not allow IMM for this calculation either, there are few CCR related concerns about the standardised floor. Relative to our decades-old CEM, we felt that SA-CCR was a significant step up in risk sensitivity and the better recognition of margining and netting would improve risk management incentives and outcomes. The limitations of CEM and the enhanced reporting framework were also key considerations for us in choosing to go ahead with implementing SA-CCR with a July 2019 start date.
While we consider the finalisation of SA-CCR an important step forwards, we also appreciate that it is not perfect in its risk capture (no standardised approach ever could be) and it does have its conservative elements (as one would expect with any standardised approach). We understand ISDA’s concerns about excessive conservatism and if there are any changes at the Basel level then we will certainly take these into consideration at the domestic level. For Banks using the Adjusted CEM, we are also considering a further carve-out of smaller ADIs from the CCR rules. Many of these smaller entities have only minimal counterparty credit risk exposure and this carve out approach is part of our broader initiative to reduce regulatory burden for the large number of smaller ADIs that we supervise.
Looking ahead a little way into the future, market participants should expect to see an updated policy framework for interest rate risk in the banking book. APRA is updating our Pillar 1 Interest Rate Risk in the Banking Book (IRRBB) approach. IRRBB in Australia is an internal model approach that applies only to the larger banks. We have had an initial consultation and, consistent with the broader trend I have alluded to already, some of the proposed changes reflect a shift towards greater standardisation and increased restriction on the use of internal models. Larger banks should also expect to see updated reporting forms and Pillar 3 disclosures which broadly reflect the standardised shocks proposed by Basel in its April 2016 paper. We are currently working through the detail and intend to conduct a further consultation on revisions to IRRBB with an updated Prudential Standard in the first half of next year.
Of course, looming large on the horizon is the Fundamental Review of the Trading Book and I don’t suppose I can really get away speaking here today and not commenting on FRTB. FRTB has been in the works for quite some time, but we are expecting the Basel Committee to land on its final rules soon, most likely in the next few months. The Basel implementation timelines were updated back in Dec 2017 and they line up market risk, along with many of the other Basel III reforms, for a 1 January 2022 start date. Over the past year, industry representatives and participants have had the opportunity to present to the Basel Committee Market Risk Group on multiple occasions and provide formal feedback on the March 2018 consultative document. There have been many discussions on the Profit and Loss Attribution test, the non-model-able risk factors framework and numerous aspects of the Standardised Approach and Trading Book/Banking Book boundary at the Basel level. I don’t intend to repeat the technical detail today, and will simply note that significant time and effort has been spent on both sides throughout the policy process.
Locally, APRA has had numerous discussions with Australian banks through the ABA to understand their views on FRTB and we have certainly found this a useful exercise. It has been useful both in our engagement and involvement with the Basel Committee Market Risk Group but also to inform our own thinking about domestic implementation. Market risk under APS 116 generally forms a relatively small part of the overall capital requirement for Australian banks, which means that review of capital standards for market risks will often take a lower priority to standards for credit, liquidity and other risks. As a result, while APRA has been a relatively early adopter of Basel measures in areas such as liquidity, with LCR and NSFR implemented ahead of most other jurisdictions, we have taken a slower “watch and wait” approach for market risk.
Globally, the regulatory community has a very clear appreciation that the F in FRTB stands for “Fundamental” and this regulatory reform will have broad implications on strategic investments in systems, the operation of trading businesses and the structure of products and markets. It goes without saying that regulators around the world will need to have a well-thought through policy framework coupled with an appropriate implementation strategy and timeline. We are committed to working with ADIs throughout this process to achieve a good prudential outcome.
As we outlined earlier in this year, we will defer our decision on the scope and timing of any domestic implementation of the Basel III market risk framework until it has been finalised at the Basel level. As many of you would know, we have been saying this for a long time now. The framework has changed significantly since January 2016 and, while I will leave it to the panel in the next session to try and read the tea leaves on what the finalisation of FRTB may look like, this evolution would seem to vindicate our watch and wait approach. Of course, when we have a clear and final framework from the Basel Committee and we have worked through the implications for the Australian prudential framework, we will come back to industry with clear direction.
Finally, I wanted to spend a bit of time on CVA risk. It has been perhaps a little in the shadow of FRTB and SA-CCR but it is an area which we consider is just as important to get right.
While credit spreads have been coming in for a long time and market volatility has been relatively low, credit spreads and volatilities can blow out quickly in times of systemic market stress. CVA risk is something of a wrong-way risk: it is one of those risks which tends to go bad when everything else goes bad. There is little diversification benefit in times of stress and during the financial crisis, we saw some very large CVA losses both domestically and internationally. In our recent 2018 ADI industry stress tests, losses from changes in CVA for some banks formed a big part of their Year 1 losses and were much bigger than losses from their trading book positions.
Despite the risks, based on our recent benchmark reviews of major bank practices in Australia, we have come to the conclusion that it is a risk which is not as well measured and managed as it could be. We have pointed out to many banks that CVA is non-linear and relying too heavily on simple CVA sensitivities to measure CVA risk is not good practice. It is worse when you are actively hedging those sensitivities and not appropriately measuring the remaining gamma and cross gamma risks. More broadly, we think CVA and XVA stress testing is something that needs to improve across the industry.
On the policy front for CVA risk, the new framework with the two options for a basic approach (BA-CVA) and a standardised approach (SA-CVA) was finalised as part of the Basel III finalisation package. The option for an internal CVA risk model was removed. However, both approaches will need to be recalibrated in light of changes in the risk weights for the FRTB Standardised approach. From a Basel perspective, this work on CVA risk will be conducted over 2019 after the FRTB framework has been finalised.
At present, APRA only allows banks to use the standardised formula for CVA risk and has not allowed the use of any internal models. The BA-CVA shares some similarities with the current formula but it does also have a few improvements. At this point in time, we have not decided whether to allow Australian banks to use SA-CVA which relies on banks internal sensitivities calculated from their CVA models. We will continue to review banks’ practices for CVA and CVA risk measurement as we make this decision. This will be one of a number policy questions that we will be consulting on.
I have sketched out some of APRA’s more recent policy initiatives affecting markets businesses and also looked a little ahead on some of the regulatory changes still on their way from Basel and APRA that might be of most interest to members at this conference. There are still complex regulatory changes ahead for markets businesses and APRA is committed to approaching these changes in a considered way and consulting effectively so that we end up with a robust and prudent regulatory framework that will put the industry in good stead over the coming years.
Thank you again for the opportunity to speak and I would be happy to take a few questions in the remaining time.