2022-04-05-IMF-The_Global_Bank_Stress_Test_40页_6mb

Summary of The Global Bank Stress Test

Core Content

The Global Bank Stress Test (GST) is a framework developed by the IMF staff to assess the resilience of banks in advanced economies (AEs) and emerging market economies (EMs) to global shocks. It is designed to provide a consistent, cross-country analysis of how macro-financial variables affect bank capital and balance sheets, incorporating cross-border spillovers and synergizing with the IMF's bilateral and multilateral surveillance activities.

The GST was introduced in the October 2020 Global Financial Stability Report (GFSR) and is based on a solvency stress testing methodology. It uses publicly available data from Fitch Connect and bank financial reports to analyze the potential impact of global shocks on bank capital, including the pandemic and other severe scenarios. The methodology involves projecting changes in capital ratios, income statement components, risk-weighted assets, and other comprehensive income (OCI) based on macro-financial scenarios and econometric models.

Main Viewpoints

• The GST is a global, cross-country framework that allows for consistent risk analysis across different economies.
• It leverages macro-financial scenarios from the IMF's World Economic Outlook (WEO) and uses country-specific panel data and econometric models to estimate the effects of these scenarios on bank capital.
• The baseline and adverse scenarios are used to project the impact of global shocks over a three-year horizon (2022–2024).
• The adverse scenario simulates a prolonged pandemic, considering factors like new variants, vaccine efficacy, and the pace of vaccine rollout.
• The severe adverse scenario is also analyzed to assess the most extreme potential outcomes, revealing that EMs are more exposed to capital shocks compared to AEs.
• The GST includes a unique global dataset of bank balance sheets, income statements, and risk metrics, spanning a 25-year period (1995–2020).

Key Information

Scope and Coverage

• The GST covers 29 major banking systems, including 24 AEs and 5 EMs, representing 70% of global banking assets.
• In each economy, the GST includes banks accounting for at least 80% of the total assets of the national banking system.
• The sample includes 53 EM banks and 204 AE banks.

Methodology

• The GST uses econometric models to project changes in P&L components, OCI, and RWAs.
• The framework is internationally consistent, drawing on the IMF Flexible System of Global Models (FSGM) for scenario design.
• It incorporates cross-border spillovers through financial and real economy linkages, and subsidiary capital shortfalls are considered in parent bank capital projections.
• A Bayesian Model Averaging (BMA) approach is used to account for model uncertainty, with over 5,000 models estimated for each dependent variable.
• The loan-loss model is linked to structural models that decompose loss rates into PDs and LGDs using a variant of the Frye-Jacobs (2012) methodology.

Findings

• AE banks have shown greater resilience, with capital ratios increasing by about 0.8 percentage points in 2020.
• EM banks face larger downside risks, are more sensitive to shocks, and have lower capital buffers.
• The GST results indicate that EMs are more exposed to severe adverse scenarios, with a higher potential for capital shortfalls.
• The capital ratio for AEs has increased significantly since the GFC, reaching a median of 17.3% by the end of 2020, up from 16.5% in 2019.
• The pandemic response has included fiscal support, monetary easing, and regulatory measures, which have helped stabilize bank capital.

Limitations and Considerations

• The use of public data limits the granularity and comprehensiveness of the analysis.
• The GST does not capture borrower-specific support policies, as such data are not available in the public domain.
• The analysis of spillovers and spillbacks is implicit, based on multilaterally consistent global models.
• The results should be interpreted cautiously, especially when compared to supervisory stress tests based on more detailed data.

Conclusion

The Global Bank Stress Test provides a flexible and rigorous framework for assessing the impact of global shocks on banking systems. It highlights the relatively lower resilience of EM banks compared to AEs, especially in severe adverse scenarios. While the GST is not a substitute for more detailed stress testing, it serves as a benchmark for national authorities and supports multilateral and bilateral surveillance. The framework is designed to evolve with new data and scenarios, ensuring relevance and adaptability in a rapidly changing global environment.