2013年-IMF国际货币组织全球_Rules_of_Thumb_for_Bank_Solvency_Stress_Testing_67页_777kb

Summary of Rules of Thumb for Bank Solvency Stress Testing

Core Content

This working paper by Daniel C. Hardy and Christian Schmieder presents a set of "rules of thumb" to assist in the design and interpretation of bank solvency stress tests. These rules are intended to provide quick, robust, and interpretable estimates of how bank capital ratios and key drivers—such as credit losses, income, credit growth, and risk-weighted assets (RWA)—behave under different levels of stress in both advanced economies (ACs) and emerging market economies (EMs). The paper emphasizes the importance of understanding the non-linear and disproportionate effects of shocks on bank solvency, particularly in severe or extreme stress scenarios.

Main Viewpoints

• Rules of Thumb as Useful Tools: Rules of thumb are proposed as a practical guide for stress testing, especially in countries with limited data or experience. They are meant to complement detailed analysis rather than replace it.
• Non-linear Responses to Shocks: Bank capital ratios and other solvency indicators exhibit disproportionate responses to large shocks. For example, severe credit losses can be many times higher than in normal times.
• Different Behavior Between ACs and EMs: The paper highlights that EMs typically experience higher and more variable credit losses, default rates, and GDP growth fluctuations compared to ACs, due to structural and macroeconomic differences.
• Importance of Regulatory Capital Measurement: The paper notes that the standardized approach (StA) under Basel II tends to be less sensitive to shocks compared to the Internal Ratings-Based (IRB) approach, which reacts more quickly to changes in risk.
• Need for Simplicity and Intuitiveness: A desirable rule of thumb should be simple, intuitive, and supported by a wide range of empirical evidence to ensure robustness and applicability across different countries and stress scenarios.

Key Information

I. Introduction

• Stress testing is a critical tool for assessing financial sector vulnerabilities and informing policy decisions.
• Rules of thumb are proposed to help in the practical application of stress testing, especially in data-scarce environments.
• The rules aim to provide general guidelines for understanding worst-case scenarios and assessing the impact of different stress levels on bank capital.

II. Methodology and Sources

• The capital ratio is the primary metric in bank solvency stress testing.
• The paper uses two main datasets:
  • Long-sample evidence: Moody's data on corporate default rates over 90 years (1920–2012).
  • Recent cross-country evidence: Bankscope data covering over 16,000 banks in 200 countries from 1996 to 2011.
• The paper accounts for differences in regulatory capital measurement, particularly between StA and IRB approaches, and highlights how these affect sensitivity to shocks.
• The projected capital ratio is calculated using the following formula:

$$
\text {Projected Capital Ratio} _ {t + 1} = \frac {\text {Initial Capital} _ {t} + \text {Projected Retained Net Profit} _ {t + 1}}{\text {Initial RWA} _ {t} + \text {Projected Change in RWA} _ {t + 1}}
$$

• The leverage ratio is also discussed, with a similar formula used for its projection.

III. Typical Banking Crises and Descriptive Rules of Thumb

A. Literature on Banking Crises

• A large body of literature exists on banking crises, focusing on macroeconomic precursors, effects, and policy responses.
• Some studies provide quantitative evidence relevant to stress testing, such as Čihák and Schaeck (2007), who analyzed 51 banking crisis episodes from 1994 to 2004 across different regions.

B. Historical Evidence on Banking Crises

• Moody's data shows that corporate default rates peaked five times over the last 90 years, with the highest peak during the Great Depression of the 1930s.
• Historical U.S. data suggests that default rates were even higher before 1900, possibly due to less developed financial systems and more volatile economic conditions.
• These historical insights support the hypothesis that EMs experience higher and more variable default rates and credit losses compared to ACs.

C. Descriptive Rules of Thumb

• The paper provides rules of thumb for credit loss rates under different stress levels:

  • Normal conditions: Median credit loss rate of 0.3% for ACs, 0.5% for EMs, and 0.6% for LICs.
  • Moderate stress: Credit loss rates increase to 1.5% for ACs, 2.5% for EMs, and 3.0% for LICs.
  • Medium stress: Credit loss rates rise to 3.0% for ACs, 4.0% for EMs, and 5.0% for LICs.
  • Severe stress: Credit loss rates jump to 5.0% for ACs, 7.0% for EMs, and 8.0% for LICs.
  • Extreme stress: Credit loss rates can reach up to 10% for ACs, 15% for EMs, and 20% for LICs.

• These rules are based on historical data and aim to provide a basis for estimating the potential impact of different stress scenarios on bank solvency.

• The paper also discusses the non-linear behavior of banks during crises, including the impact on pre-impairment income, credit growth, and risk weights.

IV. Rules of Thumb for Satellite Models

• Rules of thumb are also proposed for satellite models that link key solvency drivers (credit losses, pre-impairment income, credit growth, and RWA) to GDP changes.
• These rules are used to simulate the evolution of capital ratios under various stress conditions.
• The paper highlights the importance of using these rules to understand how shocks affect banks and to evaluate the resilience of different capitalization levels.

V. Worked Examples

• The paper provides worked examples based on the features of stylized banks in ACs and EMs.
• These examples illustrate how the rules of thumb can be applied in practice, including the simulation of capital ratio evolution during stress periods.
• The paper also discusses the importance of checking the plausibility of results using rules of thumb and highlights the role of dividend payouts and taxation in capital projections.

VI. Conclusion

• Rules of thumb are useful for assessing the impact of stress on bank solvency, especially in the absence of detailed data.
• They help in understanding the behavior of key solvency indicators and in evaluating the need for different levels of capital under various stress scenarios.
• The paper concludes that while rules of thumb are not a substitute for detailed analysis, they are an essential tool for quick and robust stress testing, especially in the context of regulatory capital requirements and policy decisions.

Appendix Overview

• Appendix 1: Summary of data used, including Bankscope data and Moody's default rates.
• Appendix 2: Supplementary evidence on the behavior of solvency parameters during crises.
• Appendix Tables: Include details on the data sources, sample sizes, and key variables.
• Appendix Figures: Provide visual summaries of the data, including the evolution of NPL ratios, default rates, and RWA during stress periods.

Tables and Figures

• Table 1: Breakdown of NPL stock ratio changes by region during crises.
• Table 2: Summary of typical credit loss levels under different stress scenarios.
• Figure 1: Illustration of NPL ratio evolution around a crisis.
• Figure 2: Historical annual default rates for all rating grades.
• Figure 3: Historical corporate bond default rates (1866–2008).

These tools are designed to be used in conjunction with detailed models and are intended to enhance the understanding and interpretation of stress test results.