EU Stress Testing Regulatory Update

‘Stress testing must involve identifying possible events or future changes in economic conditions that could have unfavourable effects on a bank’s credit exposures and assessment of the bank’s ability to withstand such changes.’
(BCBS128, art. 434¹)

Regulatory requirements to report capital plans under a set of stress scenarios are increasing. As such, financial institutions are required to adapt their internal organisation toward enterprise-wide risk management and capital planning, which involves advanced modelling, data management, and reporting tools.

Granularity, consistency, and communication across departments are the key stress testing challenges that financial institutions face in the coming years. Success is contingent upon overcoming these challenges whilst involving senior management.

This process is not without difficulties, as banks must overcome numerous hurdles, such as:

• Gathering the correct data at the transaction level
• Developing the right models to translate macroeconomic scenarios into risk parameters
• Aggregating and reporting the results of the stress testing exercise
• Transforming quantitative results into concrete short-term corrective actions to help senior managers make more informed decisions

As banks adopt and support regulatory stress testing exercises, the benefits of aligning their business needs with the regulators’ requirements become clear. For instance, some large banks currently in the process of adapting their processes to meet these requirements are also redefining their approach to stress tests. Up until recently, bank operating models allocated part-time resources to various sections of a stress test exercise. Typically, these resources maintained reporting affiliations to different divisions, teaming up only when a (regulatory) stress test cycle was required. Now, banks are building dedicated teams to cover all aspects of stress testing with the goal of developing a lean, automated, and common set of tools and processes.

Banks will still have to address what they see as the most difficult steps in developing quality idiosyncratic stress tests:

• Defining adequate scenarios
• Accessing the right data
• Properly modelling correlations with risk factors
• Effectively reporting the results

It appears that the UK Prudential Regulation Authority (PRA) is keen to follow the practices of the US Federal Reserve. By the end of this year, UK regulators are expected to ask systemically important financial institutions (SIFIs) to send regulatory reports in the XML format with granular risk data and provision information on their portfolios. This policy is called the Firm Data Submission Framework (FDSF). PRA will use the data for two purposes: to perform its own stress test and a risk assessment of an individual financial institution and to conduct a global systemic risk analysis of the UK banking sector.

Once a year, each institution will be asked to send ‘projections reports’ to the regulator, including a five-year forecast, in accordance with various stress scenarios. PRA will then compare the results of each institution to its own assessment and challenge any significant discrepancies. It is expected that institutions will have a year to explain the differences between last year’s forecast and the actual results.

This initiative requires that these institutions have data gathering platforms to pull the data with the required granularity and to report it in the XML format. Also, institutions will need to make their stress testing framework more robust to ensure they are validated by supervisors.

The Bank of England Financial Policy Committee has also recommended that the stress testing initiative be extended to the whole UK banking system. ‘Looking to 2014 and beyond, the Bank and PRA should develop proposals for regular stress testing of the UK banking system. The purpose of those tests would be to assess the system’s capital adequacy. The framework should be able to accommodate any judgements by the Committee on emerging threats to financial stability’.²

The stress tests are forward-looking and ‘will play an important part in the PRA’s judgements about a firm’s financial soundness... Stress tests will not be ‘pass/fail’ exercises, but will instead be used to assess the balance of risks arising’.³

In the European Union (EU), major banks will first have to support an Asset Quality Review (AQR) toward the end of the year.⁴ Although this initiative would typically be led by national regulators, the EBA’s aim is to harmonise methodologies, practices, and communications around these exercises. The purpose of this initiative is to classify and valuate assets held by banks to ‘dispel concerns over the deterioration of asset quality due to macroeconomic conditions in Europe’.

The timeline of this initiative is highly correlated with European Central Bank’s (ECB) Single Supervisory Mechanism (SSM) role. One of the consequences of this AQR initiative is to postpone the EU-wide stress test to 2014. However, to ensure consistency with previous years, the EBA will still provide information on the actual exposures of EU banks. For more information about the AQR, read the article later in this publication, titled: ‘Asset Quality Review: Setting the Foundation for a Standard Stress Testing Framework’. Many uncertainties remain in the EBA’s 2014 stress test. First, the capital definition of Core Tier 1 could still be the Basel 2.5 risk-weighted asset (RWA) (Tier 1 excluding hybrids) versus a Capital Requirements Directives (CRD) IV definition. Also, it is not clear if the stress testing of regulatory capital will be based on Basel 2.5, Basel III transitional, or Basel III fully phased. Therefore, it is recommended that banks create flexible regulatory capital platforms. Finally, liquidity risk may now be part of the stress testing framework, which of course provides more comprehensive results but also requires more enterprise-wide platforms.

The EBA may request various credit risk metrics sensitivity analyses (not a pass/fail exercise). The EBA may also introduce ceilings and floors in order to ensure consistency across modelling approaches and jurisdictions.

As seen in the US with the CCAR initiative, EU supervisors will now require that financial institutions produce forward-looking stress tests, taking into account all types of risks and forecasts for their portfolios. This will require tremendous organisational changes, as these stress testing exercises will involve the risk management, treasury, and capital planning departments. One solution is that banks have a dedicated department working on stress testing, which could then challenge the risk departments on their systems and models. This stress testing department would act as an internal auditor.

• Data management: Financial institutions will need to centralise all the data necessary to support stress testing models, as well as regulatory and internal reporting requirements. This system should also be able to reconcile its data with production systems to ensure result consistency. Therefore, this approach requires an enterprise-wide datamart oriented on risk and financial management. The datamart would be the data source for all risk engines, capital planning, and stress testing tools.
• Models: Internal models will be increasingly challenged by supervisors in the stress testing exercises. Banks will need to be able to document and maintain their bottom-up and/or top-down models and be consistent over time.
• Reporting: Supervisors will require more and more frequent stress testing and reporting. Financial institutions are required to publish monthly, quarterly, and annual reports in the required format. This in turn requires an automated reporting tool that can produce regulatory reports efficiently, but be flexible enough to audit and adjust those reports. Reconciliation between reports will be paramount. The reporting tools should also be able to keep pace with the regulations. There have been, for instance, several updates per year across jurisdictions (e.g., CCAR in the US) and asset classes.
• Stress testing automation: Financial institutions will need software to coordinate and centralise the stress testing process to keep consistent scenario and modelling assumptions across the balance sheet, as well as deploy and maintain a large quantity of models (e.g., periodic recalibration). The stress tests should be automated so banks can run more scenarios (e.g., business-specific scenarios). Finally, properly controlled expert judgment should generally be allowed to overwrite models on specific counterparties when real-life conditions require. Thus, stress testing automation should manage users, through workflow, auditing, and tracking.

Initially planned for this past summer, the new EBA EU-wide stress test has been postponed to allow for an Asset Quality Review to be launched for all banks being transferred to the ECB supervision mid-2014.

This requirement will most certainly reinforce the usage of stress tests by regulators, and will likely push European banks to develop a more robust framework (quantitative and qualitative), as currently seen in the US following the implementation of the Dodd-Frank Annual Stress Test (DFAST) and the third occurrence of the Comprehensive Capital Analysis Review (CCAR).

These ongoing changes represent an increasing challenge for European banks, already overwhelmed by waves of new regulations.