EBA欧洲银行-Session-II-Dale-Gray_39页_2mb

Summary of "Macro-Financial Feedbacks in Stress Testing"

Core Content

This document explores the integration of macro-financial feedback loops into stress testing frameworks, emphasizing the importance of understanding how financial sector risks interact with macroeconomic conditions and vice versa. It outlines three main approaches: the credit channel, agent-based modeling, and contingent claims analysis (CCA). These methods aim to enhance the accuracy and comprehensiveness of stress tests by incorporating behavioral responses, systemic risk, and dynamic interactions between banks, sovereigns, and the economy.

Main Points

1. Macro-financial Feedback Loops through the Credit Channel

• Traditional Stress Testing Limitations: Conventional one-round stress tests often fail to capture key macro-financial feedback loops, potentially leading to an underestimation of capital losses and systemic risk.
• Three Building Blocks:
  1. Macro Block: Models macroeconomic variables (e.g., GDP, credit, etc.) and their evolution over time.
  2. Profit and Loss Block: Estimates credit and other losses based on macroeconomic conditions and bank-specific factors.
  3. Lending Block: Integrates the impact of macroeconomic variables on bank lending behavior.
• Algorithmic Approach:
  • Step 1: Use the "macro" block to estimate $y_1$.
  • Step 2: Use the "profit and loss" block to calculate credit losses.
  • Step 3: Use the "lending" block to estimate lending changes.
  • Step 4: Calculate bank capital ratios using the formula $k_{i,1} = \frac{\text{Capital}{i,1}}{\text{RW A}{i,1}}$.
• Key Findings:
  • GDP shocks are endogenous to banks' reactions.
  • Even with capital recovery, real effects can be permanent.
  • Bank recapitalization peaks at 5% of nominal GDP.
  • Over 5 years, cumulative real GDP declines by 8% relative to baseline.

2. Structural Approach using Agent-Based Modeling

• Author: Laura Valderrama
• Key Features:
  • Incorporates heterogeneous agents (banks, noise traders, investors) and their interactions.
  • Accounts for non-linear dynamics and time-varying behavior.
  • Endogenizes key financial variables such as funding access, fire sales, and capital injections.
  • Allows for the assessment of unintended regulatory consequences.
• System Interactions:
  • A market shock (e.g., redemptions from noise traders) can morph into a liquidity shock and a credit shock.
  • These feedbacks increase default risk and slow economic growth.
  • Over 5 years, cumulative real GDP declines by 1% relative to baseline.
• Model Components:
  • Banks: Manage balance sheets, subject to Basel III regulations and market constraints.
  • Noise Traders: Have stochastic demand and are impacted by liquidity shocks.
  • Investors: Respond to bank performance and provide funding based on leverage.
• Applications:
  • The model is useful for policy simulations, including macroprudential policy and banking sector structure analysis.
  • It captures the joint dynamics of solvency, liquidity, and their impact on the real economy.

3. Contingent Claims Analysis (CCA)

• Core Concept: CCA models bank solvency based on risk-adjusted balance sheets, treating assets and liabilities as contingent claims.
• Risk Indicators:
  • Expected Default Frequency (EDF): Measures the probability of default for banks and corporates.
  • Fair-Value Credit Default Swap (FVCDS): Reflects the market's view of systemic risk and government contingent liabilities.
  • Expected Loss Ratio: Calculated as EDF × LGD and represents the present value of default risk.
• Key Applications:
  • CCA is used to estimate market-implied systemic risk.
  • It helps in understanding the relationship between EDF, FVCDS, and expected losses.
  • CCA models are used in FSAP stress tests for various countries, including the US, UK, Sweden, Germany, Netherlands, Israel, Spain, and Hong Kong.
• CCA GVAR Framework:
  • Combines CCA with Global VAR (GVAR) models to analyze macro-financial interactions across countries.
  • Projects EDFs and MCARs for different scenarios using a macro factor model.
  • Provides insights into the joint losses of the banking system and government contingent liabilities.

Key Information

• Document Focus: Enhancing stress testing by incorporating macro-financial feedbacks.
• Approaches Covered:
  • Credit Channel: Traditional, behavioral, and integrated macro models.
  • Agent-Based Modeling: Simulates interactions among banks, investors, and noise traders.
  • Contingent Claims Analysis: Estimates systemic risk and its impact on the economy.
• Outcomes:
  • Stress tests should reflect the dynamic and non-linear nature of financial systems.
  • Macro variables like GDP and credit growth are influenced by banking sector behavior.
  • Systemic risk can have long-lasting effects on economic activity.
• Policy Implications:
  • Macroprudential policies and regulatory constraints must be considered in stress testing.
  • Agent-based models and CCA provide tools for simulating and analyzing complex interactions.

References

• Part I:
  • Catalán, M. and A. Hoffmaister (forthcoming 2017), "Bank Capital and Lending: An Extended Framework and Evidence of Nonlinearity"
  • Catalán, M. and T. Xu (forthcoming 2017), "Macro-financial Feedback Loops through the Credit Channel"
• Part II:
  • Basel Committee on Banking Supervision (2015), "Making supervisory stress tests more macroprudential"
  • Bookstaber, R. (2012), "Using Agent-Based Models for Analyzing Threats to Financial Stability"
  • Beinhocker, E. (2012), "Introduction to Project CRISIS"
  • Thurner, S., Farmer, D., and Geanakoplos, J. (2012), "Leverage causes fat tails and clustered volatility"
• Part III:
  • Gray, D. and A. Jobst (2013), "Systemic Contingent Claims Analysis - Estimating Market-Implied Systemic Risk"
  • Gray, D., M. Gross, J. Paredes, M. Sydow (2013), "Modeling Banking, Sovereign, and Macro Risk in CCA Global VAR"
  • Merton, R. C. et al. (2013), "On a New Approach for Analyzing and Managing Macrofinancial Risks"
  • Gray, D. and S. Malone (2008), Macrofinancial Risk Analysis
  • Gray, D. and S. Malone (2012), "Sovereign and Financial Sector Risk: Measurement and Interactions"
  • Gray, D. and A. Jobst (2011), "Modeling Systemic Financial Sector and Sovereign Risk"

Conclusion

This document presents a comprehensive analysis of macro-financial feedbacks in stress testing, highlighting the need for more sophisticated models that account for the interplay between financial institutions and the broader economy. The three approaches—credit channel, agent-based modeling, and CCA—each offer unique insights and methodologies to better understand and model these complex interactions. The findings underscore the importance of integrating behavioral responses, systemic risk, and macroeconomic factors to produce more accurate and robust stress test outcomes.