By Simon Goldsmith, Head of Risk Solutions, SAS UK & Ireland
In December, UK bank executives breathed a sigh of relief after they all passed the annual Bank of England stress tests. It’s reassuring for consumers that financial organisations are far more able to survive another crisis now that they have: slimmed down operations, increased capital safety margins and pre-prepared contingency plans.
However, bank stress-testing teams still face two major challenges:
- IFRS 9, the new accounting standard for calculating loan impairments, requires multi-year forecasts of individual loans under different scenarios. This is very similar to a stress-testing exercise, so regulators and auditors are looking for a consistent approach with explainable differences. At the moment, embryo IFRS9 teams (who will first report in 2019) and stress-testing teams have very different methodologies and technologies. Plans for alignment are urgently needed.
- Qualitative reviews. The Bank of England (BoE) still doesn’t completely trust banks’ stress-testing analyses and made this very clear in the Qualitative Review section of the 2015 results. Banks must address these qualitative aspects, or risk being publically “failed”—just as the Federal Reserve System is doing in the US. The consequences of these qualitative failures are potentially identical to a capital-margin failure—i.e., prevention of payment of dividends, withholding of senior executive bonuses and public censure.
This article explores these challenges further and looks into how new and emerging technologies can help.
Temperature of regulatory environment
Headlines might suggest that, for the first time since the crisis, banks can demonstrate robust five-year business plans that can withstand stormy economic seas. So far, so good. Even Mark Carney, governor of the Bank of England, has claimed that banks are now “significantly more resilient than before the global financial crisis [and that the] post-crisis period is over”.
Yet, even though all seven banks successfully passed the test this time around, there is still a lot more to be done than simply ticking boxes. This sentiment was echoed in the stress-test results announcement when it was suggested that several banks still need “to make considerable improvements, including implementing and embedding model management policies more fully”. As we’ve seen, Royal Bank of Scotland and Standard Chartered were called out by the BoE as the weakest, even though they still passed. Both were found not to have enough capital strength in the stress-testing exercise. Fortunately both had already taken steps to raise additional capital before the BoE published the results.
Heightened expectations for stress testing and scenario analysis make it necessary for banks to look at risk management as not just a one-time cram session, but a year-long exercise. Being able to quickly and effectively handle any regulatory or economic scenario is essential. The current heavily manual approaches make this difficult and expensive. These expectations are not going to go away. In fact, stress testing is likely to be even stricter next year, with a much higher capital bar. That could mean a big name in hot water next time around.
Added to the challenges of regulatory stress testing come the new multi-year forecasting demands of IFRS 9. While the accounting standard was finalised in July 2014, the detailed working practices are still under debate. Recent discussions between banks, the International Accounting Standards Board (IASB) and auditors suggested an approach that would add considerable complexity to impairment forecasting calculations.
Challenges for stress-testing and IFRS 9 processes
Recent research by the Global Association of Risk Professionals (GARP) has shown that, while progress has clearly been made and banks have improved their approaches to stress testing, the technology infrastructures remain old and complex. In addition, most are not yet considering how they could exploit new IFRS 9 infrastructure to enhance stress-testing approaches. This limits their ability to reap the business benefits of the latest forecasting systems, i.e.,
- accelerated run times,
- much lower costs of operation,
- flexibility to deliver new scenarios and
- tight process governance and control.
This approach does not appear to be the result of budget freezes or cuts, as most financial institutions have significantly increased spending on stress testing. However, most of this investment has either gone into tactical extensions of existing systems or the creation of additional manual processes. We regularly hear stories of demoralised employees working late into the night across many months of the year simply because too many compliance gaps have been closed with resource-wasteful manual processes.
On top of this, teams will need to factor in the additional workload of IFRS 9 forecasting, which could be significant. Put simply, yet more manual processes and tactical extensions won’t work, and it certainly won’t close many of the regulatory governance gaps. Current systems are intrinsically too prone to error, slow and inflexible.
Key areas for future changes in stress-testing architectures
The GARP report was based on a survey in spring 2015 and highlighted four areas of readiness for consideration:
- Managing data. Banks require comprehensive data-management capabilities that include data-quality, data-lineage and metadata documentation in a transparent and readily searchable form through the entire stress-testing life cycle.
- Monitoring model risk and performance. Banks require capabilities to support model development as well as model governance, enabling independent review and validation of models used in internal capital planning.
- Model-execution platforms need to perform huge numbers of calculations, while executing a variety of models on large amounts of granular-level data. These platforms must also aggregate the results to any desired level for analysis and reporting, while including process-management capabilities that ensure all necessary steps are completed, monitored and repeatable.
- Banks need to specify scenarios and consolidate modeling results into balance sheets, financial statements and capital plans while orchestrating the various aspects of the stress-testing process and consolidating the results from various systems.
To GARP’s analysis from the summer we would add:
IFRS 9 alignment. IFRS 9 and stress-testing teams need to develop and exploit a common strategic architecture. Infrastructure is likely to be developed initially for IFRS 9, then extended to provide the credit-forecasting elements of stress testing.
Emerging approach being taken by leading banks
Several large banks are now developing new strategic architectures to meet both IFRS 9 and stress-testing requirements. These architectures use combinations of:
- sophisticated model-development factories to efficiently develop, validate and manage sometimes hundreds of interconnected models,
- multi-model execution engines to rapidly generate detailed, multi-scenario forecasts and
- tightly governed consolidation and reporting platforms for final review and forecast signoffs.
Not only does this enable them to comply with today’s regulator and auditor needs, but it also provides a flexible platform to cater for inevitable changes to business, regulation and external reporting.
Banks, therefore, need to seize the opportunity to fully understand the ever-evolving state of their business and the new technology options now available. Deploying such technologies not only simplifies compliance but also releases budgets. Banks can then address what many CEOs consider bigger long-term business issues—digitising their entire offerings to improve services to customers and therefore compete with new entrants to the banking sector.
Find out more about how technology can support banks with IFRS 9 compliance.
Photo Attribution: © Depositphotos/tupungato