EBA欧洲银行-BSLoss-A-comprehensive-measure-for-interconnectedness-K.-Fink2C20U-Kruger2C20B.-Meller2C20L.H.-Wong_28页_1mb

Summary of $BSLoss$ - A Comprehensive Measure for Interconnectedness

Core Content

This paper introduces $BSLoss$, a novel measure of interconnectedness in the banking system, which quantifies the increase in expected credit losses due to contagion risk. The measure is designed to reflect the economic impact of interbank credit networks and is based on a multiple-round contagion algorithm that simulates how a deterioration in credit quality of one or more banks can propagate through the system and affect other banks.

Main Viewpoints

1. Interconnectedness and Financial Stability

• Interbank credits have two-sided effects: they can improve risk sharing and diversification, but also expose the system to contagion risk.
• Contagion risk refers to the spillover effects of a bank or group of banks' distress, which can lead to default or further financial instability.
• The paper emphasizes the importance of contagion risk in the context of systemic risk and regulatory reform, particularly after the 2008 financial crisis.

2. Limitations of Existing Measures

• Default cascade models are useful for understanding contagion but are backward-looking and not responsive to small credit deteriorations.
• Centrality measures are forward-looking and more responsive, but are difficult to interpret in economic terms and lack empirical grounding.
• These limitations motivate the development of $BSLoss$, which combines the strengths of both approaches.

3. The $BSLoss$ Model

• $BSLoss$ is based on the transmission of credit risk through interbank networks.
• It models the contagion process as a recursive mechanism, where the distress of one bank affects its creditors, who in turn affect their creditors, and so on.
• The model uses empirical relationships between Tier 1 capital ratios and probability of default (PD) to simulate the spread of distress.
• It accounts for both expected and unexpected credit losses, which are reflected in asset devaluation and increased regulatory capital requirements.

Key Information

1. Methodology Overview

• The model simulates the impact of a change in a debtor bank's PD on the creditor bank's Tier 1 capital ratio.
• The Tier 1 capital ratio is updated based on changes in risk-weighted assets (RWA) and loss given default (LGD).
• The algorithm iteratively computes the PD for all banks over multiple rounds, considering the recursive nature of the banking network.
• The contagion effect is modeled using a logit-regression that estimates the marginal effect of changes in the Tier 1 capital ratio on PD.

2. Algorithm Description

• The algorithm starts with an initial PD at $t = 0$ and applies an exogenous shock at $t = 1$ to a set of banks $S$, increasing their PD by a parameter $\varphi$.
• For $t \geq 2$, the PDs of all banks are updated based on the changes in their Tier 1 capital ratio and the PDs of their counterparties.
• The termination condition is when the change in PD between iterations is below a small threshold $\epsilon$, indicating that the system has reached a stable state.
• The $BSLoss$ is calculated as the difference in total assets between the initial and final state of the system after the contagion process.

3. Empirical Foundations

• The model is empirically grounded, using real data on bilateral exposures and regulatory capital ratios.
• The logit-regression is used to estimate the relationship between Tier 1 capital and PD, which is essential for the quantitative analysis of the contagion effect.
• The regression includes control variables such as profitability, liquidity, and management quality, which help isolate the effect of capital on default risk.

4. Policy Implications

• The model can be used to evaluate the effectiveness of systemic capital buffers in mitigating contagion risk.
• It allows for policy experiments, such as increasing Tier 1 capital for highly interconnected banks, to assess how much systemic losses can be reduced.
• The model is flexible, and can be extended to analyze shocks from outside the banking system, such as sector-specific shocks.

Conclusion

• $BSLoss$ provides a comprehensive and economically interpretable measure of interconnectedness in the banking system.
• It captures the recursive nature of contagion and reacts sensitively to changes in credit risk.
• The model can be used to support macroprudential policy decisions and assess the resilience of the banking system to different types of shocks.
• It offers a more accurate and forward-looking approach to measuring systemic risk than existing models like DebtRank.