EBA欧洲银行-Slides-Discussant-Klaus-Duellmann2C20ECB_9页_206kb

Summary: Specialisation in Mortgage Risk under Basel II

Core Content

This document presents a research paper discussing the impact of Basel II regulations on bank behavior, particularly in the mortgage lending sector. The study focuses on how banks respond to regulatory changes by adjusting their lending portfolios and pricing strategies. It also examines whether a specialization effect occurs, where banks may opt to focus on either high or low risk lending based on the regulatory framework.

Main Findings

• Regulatory Impact on Pricing: Lenders that adopted the more risk-sensitive Internal Ratings-Based (IRB) model after 2008 reduced the interest rates on low Loan-to-Value (LTV) mortgages by approximately 31 basis points compared to those using the Standardised Approach (SA).
• Portfolio Shift: These IRB lenders increased their portfolio share of low-LTV mortgages by an additional 11 percentage points, indicating a shift towards lower-risk lending.
• Specialization Effect: The study confirms that the specialization effect is observable in practice. IRB lenders have a higher portfolio share of low-LTV mortgages compared to SA lenders.
• Pass-Through Effect: On average, a 1 percentage point increase in risk weight corresponds to a 1 basis point increase in interest rates.
• Risk Weight Dispersion: For mortgages with LTV < 50%, the average difference in risk weights between IRB and SA lenders is 30 percentage points, leading to a 30 basis point interest rate gap.

Approach

Approach (1)

• Methodology: The paper uses granular, loan-level data to assess the impact of risk-based regulation.
• Framework: It adopts a triple difference (DDD) approach, similar to Behn, Haselmann, and Wachtel (2015).
• Scope: The analysis is centered on the UK mortgage market, but the findings are considered broadly relevant due to the significant role of mortgage lending in the European economy (approximately 4 trillion € or 23% of total loans in the Euro Area).
• Data Sources:
  • FCA Product Sales Database (PSD) at the loan level
  • Survey data providing detailed lender risk-weight information
  • Historical regulatory data from the Bank of England
• Dependent Variable: Initial interest rate

Approach (2)

• First Strategy: A triple difference estimator is used to test the hypothesis of specialization by LTV under Basel II, considering the shift from Basel I to II, IRB vs. SA banks, and high vs. low LTV thresholds.
• Robustness Checks: The results are tested for robustness against various specification assumptions and additional controls such as capital buffers and LTV band granularity.
• Second Strategy: The effect of risk weights on mortgage rates is analyzed using data from 2009 to 2015, controlling for bank-time fixed effects, LTV bands, and industry-wide variations in competition and risk.

General Remark

• Adverse Selection: The introduction of IRB models under Basel II was acknowledged as a potential source of adverse selection, where banks might choose the regulatory approach that minimizes their capital requirements.
• Regulatory Response: The Basel Committee on Banking Supervision (BCBS) recognized this issue and suggested mitigating it through supervisory processes.
• Empirical Evidence: The paper provides empirical evidence that adverse selection indeed occurs in practice, with 77% of the observed risk weight dispersion attributed to credit risk in the banking book.
• Risk Composition: Up to 75% of this dispersion is explained by differences in the risk composition of banks' assets.

Policy Implications

Policy Implications (1)

• Reforms and Capital Requirements: The observed specialization effect may explain the lower capital requirements for IRB banks, suggesting that they are not necessarily more risky but have different risk profiles.
• Future Risk Exposure: Larger IRB banks may take on higher risk in the future, which could have implications for systemic stability.
• Pro-Cyclicality: If risk weights are pro-cyclical, the incentives for specialization are stronger during economic booms than during downturns.
• Macro-Prudential Measures: These measures, such as counter-cyclical capital buffers, can amplify the specialization effect, potentially increasing concentration risk.

Policy Implications (2)

• Balancing Adverse Selection: Although adverse selection is not new and was anticipated during Basel II discussions, its implications must be balanced with the benefits of a risk-based framework.
• Capital Arbitrage Reduction: A closer alignment between regulatory and internal risk measurement reduces the scope for capital arbitrage.
• Improved Risk Management: More risk-sensitive regulatory approaches encourage better risk management practices within banks.

Technical Remarks

• Risk Weight Calculation: While the Basel II agreement allowed banks to use internal models, only certain risk components (e.g., Probability of Default, Loss Given Default) were estimated by banks. The dependence structure was hardwired into the risk weight functions, not determined by internal models.