EBA欧洲银行-Session-VI-Mario-Quagliariello_12页_948kb

EBA Stress Test Design: 5 Q&As Summary

Introduction: Are Stress Tests “Mad, Bad, and Dangerous”?

Stress tests are increasingly used by banking supervisors, both in the EU and the US, as a common tool for assessing bank resilience. However, they face challenges in terms of credibility, reliability, and accountability. The debate around stress tests involves questions of their interpretation, media influence, and the balance between cost and benefits. The introduction sets the stage by referencing the idea that stress tests may be "Mad, Bad, and Dangerous," highlighting their complexity and the need for careful design and implementation.

Core Content

• Pass/Fail Stress Tests:

  • Pass/fail stress tests are easier to interpret but may create a false sense of security.
  • The absence of clear capital thresholds makes results less intuitive but allows for a focus on impacts and vulnerabilities.
  • A pass/fail threshold is necessary in recapitalization stress tests (system-wide) and can be useful in bank-specific (P2) exercises.
  • No pass/fail threshold implies no bank can pass.

• Transparency and Self-Fulfilling Prophecies:

  • Transparency in stress test results supports market discipline and complements regulatory assessments.
  • There is a debate on whether transparency might lead to self-fulfilling prophecies, especially in pass/fail scenarios.
  • Transparency is critical in crisis stress tests for rebuilding public trust.
  • Disclosing the design, methodology, and assumptions is important for accountability, even if the details of top-down models are not always made public.

• Use of Banks' Models:

  • Stress tests use banks' internal models as part of the existing regulatory framework.
  • These models are flawed, as evidenced by their failure during the financial crisis.
  • Bottom-up approaches can be constrained with caps and floors to limit manipulation.
  • Top-down models, while not manipulable, are also subject to model risk and may reduce banks' incentives to invest in risk management.
  • Supervisors are increasingly interested in how banks generate their results, not just the numbers.
  • Models are the best available tool, but they are imperfect. Constraints, benchmarking, and granular transparency can mitigate model risk and improve credibility.

• Scenario Credibility:

  • Scenario selection is relatively standardised, typically involving 2 or 3 scenarios: a baseline and one or more adverse scenarios.
  • Multiple scenarios are difficult to manage in system-wide stress tests but are important for testing a range of risks.
  • Some risks, such as conduct risk and cyber risk, are not easily captured by standard scenarios.
  • The UK has introduced cyclical and exploratory scenarios to test more structural risks.
  • There is a concern that stress tests may be perceived as forecasts, and that they may not fully reflect real-world events.
  • Political economy considerations affect the design of stress tests, as policy makers prefer to portray the financial system as resilient.

• Risks to be Covered:

  • Stress tests should be comprehensive but not overly complex.
  • There is a trade-off between comprehensiveness and simplicity.
  • Some risk drivers have a greater impact on banks than others.
  • Stress tests may not fully capture all risks, especially those related to conduct and cyber threats.
  • The methodology should remain neutral with respect to banks' business models and focus on key risk factors.

Key Information

• Pass/Fail Thresholds: Used in system-wide recapitalization tests and as an option in bank-specific tests.
• Transparency: Enhances market discipline and accountability, but must be balanced with the risk of self-fulfilling prophecies.
• Model Risk: Banks' internal models are used but flawed; constraints and benchmarking are essential for credibility.
• Scenario Design: Standardised but limited in scope; exploratory scenarios are being considered to test structural risks.
• Risk Coverage: Comprehensive but with trade-offs; some risks are not easily captured by standard scenarios.

Conclusions

• Stress tests are influenced by both practical and political economy factors, which affect their role in different economic conditions.
• Despite these challenges, stress tests remain a crucial tool for addressing disaster myopia, as they provide a way to assess bank resilience before crises occur.
• To enhance credibility, stress tests need a consolidated (if not standardised) methodology, clear data requirements, and a stable timeline.
• It is important to acknowledge their limitations and mitigate design risk, as they are not a panacea but a valuable tool for regulators.

EBA Information

• Name: European Banking Authority
• Address: Floor 46, One Canada Square, London E14 5AA
• Contact:
  • Tel: +44 207 382 1776
  • Fax: +44 207 382 1771
  • Email: info@eba.europa.eu
  • Website: www.eba.europa.eu