Banking on resilience: putting stress to the test
June 2017 | COVER STORY | BANKING & FINANCE
Financier Worldwide Magazine
June 2017 Issue
Stress testing emerged in the wake of the 2007-2008 financial crisis as a means of testing the ability of financial institutions to withstand a hypothetical adverse event. In the years since, it has been actively promoted by authorities as a prime tool for assessing the resilience of the largest and most complex banks.
Testifying to this is a 2016 KPMG report, ‘Stress testing: A benchmark analysis of systemically important financial institutions’, which contends that banks are now better able to demonstrate to regulators the quality of their approach and controls, having made significant progress over the past few years toward developing stress test frameworks. However, even with the substantial improvements that have been made to the technical application of stress testing, work remains to be done to enhance operational delivery and embed stress testing even further.
Similarly, a 2014 survey of senior banking officials in the US and Europe – undertaken by SAS Institute and Longitude Research – found that more than half believed stress testing would help them drive strategic decisions. Furthermore, a “healthy minority” characterised themselves as being highly mature in terms of their organisation’s stress testing. On the flip side, the survey notes that some banks are only at an early stage of maturity, having only established a basic framework for stress testing, while others are still outlining the resources required.
Despite this variance in professed maturity, stress testing is officially considered bona fide – an in vogue assessment tool which is naturally subject to continuous monitoring and refinement. For example, the European Union (EU)-wide stress tests carried out by the European Banking Authority (EBA) take place on a regular basis, as do those set by the Bank of England (BoE). The objective of such tests, says the EBA, is to “provide supervisors, banks and other market participants with a common analytical framework to consistently compare and assess the resilience of large EU banks to adverse economic developments”.
In terms of the 2016 EU-wide stress tests, although a number of banks were found to have particularly poor end point capital positions (Italy’s Monte dei Paschi di Siena, the UK’s Royal Bank of Scotland (RBS) and Ireland’s Allied Irish Banks among them), the overall results suggested that the top 51 banks in Europe would be strong enough to survive another financial crisis.
In summary, the EBA reported that: (i) from a starting point of 13.2 percent common equity tier one (CET1) ratio, the stress tests demonstrated the resilience of the EU banking sector to an adverse scenario, with an impact of 380 bps CET1 on average; (ii) the tests, which do not contain a pass/fail threshold (once the case but now abandoned), will be used to inform supervisors’ ongoing review of banks and guide their efforts to maintain capital in the system and support the ongoing repair of balance sheets; and (iii) alongside the results of the tests, exceptional transparency had been provided as to EU banks’ balance sheets, with over 16,000 data points per bank, a big step toward enhancing market discipline. Despite the positivity, Andrea Enria, chairman of the EBA, was quick to downplay the results, warning that the outcome of the tests “did not represent a clean bill of health” and “there remains work to do”.
In comparison, the stress test run by the BoE in 2016, while revealing the capital inadequacies of some banks (RBS being highlighted once again), reported that the UK banking sector would be resilient in the face of a major economic shock. However, like the EU-wide stress tests, the BoE results did not pass without criticism. Several senior figures, such as Sir John Vickers, once chair of the Independent Commission on Banking, a UK government inquiry into ways of reforming the UK banking sector, stated their belief that the test was not exacting enough and needed to be improved.
With banks having digested the results of the 2016 stress test, the focus now switches to the 2017 version. This year, the test requires the seven participating banks to process the new exploratory scenario – which the BOE says will consider “the impact of weak global supply growth, persistently low interest rates, and a continuation of declines in both world trade relative to GDP and cross-border banking activity” – and run a separate IFRS9 (a new standard which aims to simplify the accounting of financial instruments) exercise, in preparation for its inclusion in the 2018 stress test.
Streamlining stress – the challenges
With banks having begun 2016 in a strong position as regards their capital levels – a year which subsequently ended with the biggest average CET1 reduction in stress test history – the big question is how banks should be preparing for the 2017 test. Indeed, with most banks again well capitalised, an uptick in the severity of the test is a probability, if not a necessity.
“Banks should streamline and enhance their stress testing process so that it is sustainable and can be repeated in an efficient manner,” says Rob Smith, a banking partner at KPMG UK. “Based on our discussions with banks, significant challenges are poor data quality and lack of automation. Also, banks should perform dry runs to test the new or changed parts of the processes to avoid unnecessary issues during the stress testing cycle.” Furthermore, Mr Smith is aware that such issues are likely to be exacerbated in 2017 with the introduction of the BoE’s additional Biennial Exploratory Scenario (BES) – a process which will require the participant banks to model and document two different stress projections.
Likewise, Simon Brennan, director of regulatory strategy at Deloitte, although largely convinced that banks’ stress testing capabilities have matured since the introduction of the BoE’s stress testing exercise, is nevertheless concerned that processes in general are still resource intensive and less flexible than is ideal. He also agrees that the biggest challenge to emerge from the 2017 exercise will be managing the introduction of the BES. “No other supervisor has done anything like this before. It is uncharted waters,” he suggests.
Wholly unconvinced by the BoE stress testing regime, as well as the process itself, is Kevin Dowd, professor of finance and economics at Durham University Business School. “Banks are forced to comply with the stress tests and will doubtless view their involvement primarily as a compliance exercise,” he says. “Whether their involvement will actually lead to improved risk management is another issue and I am not convinced. Nonetheless, it suits everyone to spin the line that these exercises are actually useful.”
Stress testing – key issues
For the banks obliged to face the annual BoE stress test and its less regular EBA equivalent, there is a range of issues surrounding the processes involved, including the scope of banks under examination, the complexity of the tests, related disclosure issues and the demand for additional exercises such as BES and IFRS9.
“Many of the key problems that banks face arise from – or at least are compounded by – the complexity of the stress test exercise,” suggests Susanne Hughes, a director in Deloitte’s financial services risk advisory practice. “The timetable for running the scenarios, guidance from supervisors and data requirements all contribute to the complexity. This year in particular, banks will also find it challenging to factor in a number of regulatory, accounting and structural changes that will take place over the horizon of the forecasts.”
With most of the largest banks subject to multiple stress tests from multiple global regulators, further complexity arises from diverging practices, leading to difficulties for banks in achieving high levels of efficiency and effectiveness in compliance. “We would argue that the level of complexity in global stress testing has risen to the point where the degree of effort required by institutions is not matched by the value they derive from the exercise,” says Mr Smith. “Stress testing processes have evolved over time and often at short notice, which has resulted in the proliferation of tactical components rather than being designed to fit the size and complexity of the organisation. Further, these structures are not fully supported with formal roles and responsibilities, making it challenging to hold the supporting teams accountable for the quality and timeliness of their inputs.”
Looking ahead, the stress testing process is set to become even more complicated with the inclusion of IFRS9 forecasting in 2018 – an addition which is expected to have a substantial financial impact on banks and involve significant implementation challenges. “The key to executing a successful cycle is to have an agile, high performance environment with an associated model repository of interchangeable validated models,” advises Lee Thorpe, head of risk business solutions at SAS UK & Ireland. “This ensures that entire portfolios are utilising appropriate models that are sensitive to the key risk drivers that have been through validation and review. A centralised, top-down approach will not only allow banks to remain flexible to changing risk policies and regulations, but also effectively track model issues, challenges and remediation across the business while allowing time for controlled changes to the process.”
Enhancing strategy and governance
Among the advantages for banks of participating in a robust stress test is that the process can make management more forward looking, help identify events which could threaten viability, and act as a risk management and strategic planning tool. “The key to getting banks in better shape is better incentives on the part of those who make the big decisions,” says Professor Dowd. “These are currently weak, not least because of central bank and government policies – such as deposit insurance, lender of last resort and ‘too-big-to-fail’ – that positively encourage excessive risk taking. These are the problems to be solved. Once the right incentives are in place, the banks will automatically become stronger.”
Part of the problem, according to Mr Thorpe, is the lack of investment in data and modelling platforms, as well as enhanced governance activities, over many years. “It is generally easier to evolve the process slightly each year to keep up with the increasing demands of the regulators,” he says. “Without a process that has failed or a trigger for change, banks under severe cost pressure will not prioritise project budgets in relation to risk systems. But the inclusion of IFRS9 may be the point where current systems and process fail, and this needs to be planned now in order to have any chance of being ready in 2018.”
Another aspect of the stress testing process that banks particularly struggle with is a lack of visibility as to the effectiveness of their stress testing, i.e., they do not have key performance indicators in place – an oversight which makes it harder for them to decide where investments should be focused to improve the process. “Poor data is one of the main challenges,” confirms Mr Smith. “This further increases governance activities and time required downstream to remediate underlying data matters. Establishing key performance indicators will help to identify root causes and, on the flip side, quick short term wins to improve the process.”
Informative and insightful or non-exacting and lacking robustness? The jury is still out on what stress tests actually tell us about the strategic decisionmaking, appetite for risk and overarching resilience of the UK banking sector, or any other jurisdiction for that matter.
Many commentators, while noting the accepted worth of the 2016 BoE and EU stress tests, recognised that the calibration of testing frameworks and the principles underpinning them required significant attention. And so, with this need now apparently fulfilled, there is an expectation that the 2017 BoE stress test round will deliver a much greater insight into banks’ resilience. “In the immediate aftermath of the global financial crisis, what really mattered was banks’ ability to withstand further shocks,” says Mr Brennan. “Now that in aggregate the banking system is much more resilient, the real value from the stress test exercise comes from the insights supervisors get into banks’ risk management and governance capabilities.”
Others, though, are far less sanguine. “Whether there are improvements here or there, to hurdle ratios, for example, the stress tests themselves and the methodology underlying them are fatally flawed,” says Professor Dowd. “I do not believe that regulatory stress tests give us any true insights into the state of the banks. Instead, they are used to give the impression of resilience that is not there. The real purpose of the stress testing programme is to assure us that the banking system is sound. Whether it is or not is quite another matter.”
That said, the consensus appears to be that the health of the vast majority of UK banks is improving, with stress tests forecast to continue to play an integral part in monitoring the health of the sector. “The nature of stress testing implies that the complexity will increase each year by either altering the severity of the stress or the scenario to identify any potential weaknesses,” says Mr Thorpe. “Due to the nature of risk and how it can arise from unforeseen events, I doubt we will ever come to a situation where all banks can accurately quantify all risks under both business as usual and extreme scenarios. On this basis, it is likely we will continue to find some banks are more exposed to certain risks than others and, without the industry holding substantially more capital, stress tests will continue to have winners and losers.”
© Financier Worldwide