EBA欧洲银行-Stress-Testing-the-Credit-Risk-of-Mortgage-Loans-C.-Greve2C20L.-Hahnenstein_35页_889kb

Summary of "Stress Testing the Credit Risk of Mortgage Loans: The Relationship between Portfolio-LGD and the Loan-to-Value Distribution"

Core Content

This paper explores the relationship between the Loss Given Default (LGD) of mortgage loan portfolios and the Loan-to-Value (LTV) distribution under stressed recovery rate scenarios. The key focus is on how the shape of the LTV distribution affects the average portfolio LGD when recovery rates decline due to adverse economic conditions, such as a drop in property prices.

Main Points

• Stress Testing and Credit Risk: Stress testing has become a critical tool for banks and regulators to assess credit risk in a forward-looking manner. It is used to estimate expected loan losses and regulatory capital charges under adverse scenarios.

• LTV and LGD Relationship: The LGD of a mortgage loan is a function of both the LTV and the recovery rate (RR). The formula for LGD is:
  $$
  LGD_i = \max \left[ 0, 1 - \frac{RR_i}{LTV_i} \right]
  $$
  This shows that LGD is non-linear and asymmetric in relation to LTV and RR.

• Portfolio LGD Calculation: The mean exposure-weighted LGD of a portfolio is calculated using:
  $$
  LGD_P = \sum_{i=1}^n \frac{L_i}{\sum_{i=1}^n L_i} \cdot LGD_i
  $$
  This indicates that the average portfolio LGD is highly dependent on the distribution of LTV ratios across the portfolio.

• Impact of LTV Distribution: The paper demonstrates that the degree of sensitivity of the portfolio LGD to a stressed recovery rate depends on the dispersion of the LTV distribution. Banks with more dispersed LTV distributions experience less dramatic changes in average LGD under stress.

• Closed-Form Approximation: The authors derive a closed-form solution for the mean portfolio LGD under the assumption that the LTV distribution follows a beta distribution. This formula is intended to serve as a benchmarking tool for risk managers, rating agencies, and regulators, especially when loan-level data is not available.

• Empirical Insights: The paper references previous studies that have explored the determinants of LGD, including the LTV ratio and housing market conditions. These studies show that current LTV is often the most significant predictor of LGD, and that housing market cycles play a crucial role in explaining LGD variation.

• Real-World Application: The authors conduct a simulation study to evaluate the performance of their formula across hypothetical real-world portfolios. They simulate different LTV distributions and recovery rate stress scenarios, showing that the formula provides a reasonable approximation of the exact LGD values when full loan-level data is not available.

• Key Findings from Simulations:

  • The mean portfolio LGD increases with the dispersion of the LTV distribution.
  • The sensitivity of portfolio LGD to recovery rate stress decreases as the LTV dispersion increases.
  • When the recovery rate drops below the minimum LTV in the portfolio, the LGD becomes insensitive to further changes in the LTV distribution.
  • The formula is robust and can be used as a rule-of-thumb for regulatory and risk management purposes.

Key Information

• Stressed Recovery Rate: A drop in property prices leads to a reduction in recovery rates, which in turn increases the LGD for the portfolio.

• LTV Dispersion: The variability in LTV ratios across a portfolio significantly affects the sensitivity of the portfolio LGD to changes in recovery rates.

• Beta Distribution Assumption: The closed-form solution is based on the assumption of a beta-distributed LTV, which allows for a more tractable and generalizable approach to LGD estimation.

• Benchmarking Tool: The derived formula is proposed as a practical tool for benchmarking LGD under stress, particularly for regulators and rating agencies who lack access to detailed loan-level data.

• Limitations: The paper notes that without proper consideration of LTV distributions, comparisons of stressed LGDs across banks can be misleading or flawed.

• Policy Implications: The paper highlights the importance of macro-prudential instruments like LTV caps, which can influence the resilience of mortgage portfolios and reduce procyclicality in the banking system.

Conclusion

The study emphasizes that LTV distribution is a critical factor in determining the impact of stressed recovery rates on portfolio LGD. A proper understanding of this relationship is essential for accurate stress testing and regulatory benchmarking. The proposed closed-form solution offers a useful approximation for practical applications, especially when detailed data is not available.