2024-05-26-欧洲央行-用于财务稳定性应用的压力测试方法的进展（英）_100页_3mb

Key Advancements in Stress-Testing Methodologies for Financial Stability

Overview and Evolution

• Stress testing evolved from internal risk management tools to a key regulatory component following the 2008 crisis, with the European Central Bank's (ECB) Financial Stability Committee Working Group on Stress Testing (WGST) advancing methods between 2018 and 2022.
• Post-crisis, stress testing addresses shifting financial landscapes, incorporating policy impacts (e.g., climate change, COVID-19) and promoting trust through transparency.

Top-Down Stress Testing

• Dominated by models for credit risk (expected default rates, LGDs), market risk, and profitability (NII, NFCI, dividends).
• Credit Risk: Bayesian stochastic models (e.g., SSVS) improved scenario sensitivity; sector-specific quantile regressions captured heterogeneity during events like COVID-19.
• Market Risk: Expanded to include liquidity reserves, counterparty risk (CCR), and net trading income (NTI), with tools like EPIC and SHS-G integrating granular data.
• Validation Enhancements: Bayesian back-testing and cross-country comparisons ensured robustness.

Macro-Micro Interactions and Macroprudential Stress Testing

• The BEAST model integrates banking and macroeconomic dynamics, incorporating feedback loops (real economy-bank~sector interaction, solvency-funding feedback) to assess systemic risks.
• Applications include climate stress tests amplifying losses through contagion; impact assessments of policies like Basel III via stochastic simulations; support for ECB's monetary policy evaluations.

System-Wide Stress Testing

• ISA model evaluates contagion across banks, funds, and insurers using granular network data, capturing first-round and second-round effects.
• Initial applications in climate scenarios showed significantly amplified systemic losses compared to bank-only tests; focus on dual risk exposures in funds.

Conclusions and Future Agenda

• Stress testing in Europe remains an evolving field; advancements over the past four years enhanced resilience analysis under diverse shocks.
• Key priorities: Incorporating new datasets (e.g., EMIR, Anacredit), improving model validation, evaluating emerging risks (cyber, climate), and ensuring policy-calibrated outcomes.

Summary of Key Future Steps

• Enhanced Granularity: Refine models with transaction-level data and statistical advancements.
• Robust Validation: Develop automated ex-ante/ex-post frameworks to improve credibility.
• Communication: Support policymaker needs beyond traditional risk metrics through scenario generation.
• Emerging Risks: Focus on policies for climate and cyber threats, leveraging system-wide models.