On 22 January, the European Banking Authority (EBA) launched a public consultation on possible changes to the European Union- wide stress test, with the aim of making it more relevant and cost-effective. The proposed new framework, however, would increase complexity and hamper comparisons, a credit negative for European banks.1
The proposal envisages splitting the stress test into two parts, resulting in two separate sets of results. Both parts of the test would use the same stress scenarios. One part would be conducted by the banking supervisor (the supervisory leg following EBA guidance) and would be based on a common set of assumptions and constraints with limited variations. The other stress test undertaken by individual banks (the bank leg) would disregard constraints from a list provided by EBA, provided that this relaxation is justified. Moreover, the EBA has not ruled out the possibility of allowing banks to simply rely on their internal capital adequacy assessment process (ICAAP), an ongoing internal assessment of a bank’s capital needs.
The supervisory assessment would serve as the starting point for supervisory decisions and would be directly linked to the setting of Pillar 2 Guidance (P2G). The EBA would only provide the capital depletion under each stress scenario and its main driver and investors would therefore be able to better understand the rationale behind the level of P2G (provided that P2G is disclosed, which is still open for discussion). The bank’s assessment of the stress test would complement the supervisory leg insofar as the public disclosures will be more detailed based on a common disclosure template similar to the current one. Additionally, banks would be required to provide explanations about the difference between the capital depletion determined by the supervisors and its own assessment.
The EBA’s proposal to revise the stress test framework comes after criticism from the European Union Court of Auditors and analysts that the European Union exercise is less reliable than similar stress tests conducted in the US and UK, for example. The banks have also complained that the current framework is excessively costly and labour-intensive for all parties. They say that many of the assumptions are oversimplifications that undermine the value of the test and distort the rank ordering of banks.2
The EBA itself explicitly acknowledges in its discussion paper that the current exercise tries to achieve conflicting aims that are difficult to achieve in a single exercise. The EBA seeks to achieve four objectives: relevance, transparency, comparability and cost efficiency.
The EBA also seeks to increase and clarify ownership of the results, hence the two sets of stress test results. Last but not least, the EBA clearly states that its stress would be a micro exercise meant to underpin the supervisory work, the Supervisory Review and Evaluation Process (SREP), rather than a macro-prudential tool.
From a credit perspective, undertaking two sets of stress tests that provide two separate sets of results will not make comparisons between individual banks (or comparisons between the two results) an easy task, even if the banks explain the differences. The supervisory part of the test would apply exactly the same measures to all banks and thus would provide a comparable outcome, but disclosure would be more limited than in the past. Meanwhile, the bank part of the stress test would include more detailed information, but the methodology the banks use would be more flexible, making comparisons among banks more difficult.
Banks will likely argue that their own assessment is more realistic as it would presumably better capture the bank’s risks, which could undermine the value of the supervisors’ approach. Conversely, supervisors will argue that it is of the utmost importance to assess banks’ resilience to stress against a single set of assumptions. There is a chance the two approaches complement each other, regardless of whether or not comparisons between banks’ assessments will be meaningful. Still, investors might not like receiving two sets
of information that
they will not
be able to easily compare
– all the more so since supervisors will undertake more
limited quality assurance on
the banks’ component.
The EBA’s proposed stress test revision will not affect this year’s EBA stress test, which will use the existing framework. The banking authority is seeking feedback on its proposed changes by 30 April 2020. We note that the EBA has not ruled out keeping the current stress testing framework, should the proposed framework attract too much criticism.