2015年-IMF国际货币组织全球_The_Role_of_Bank_Capital_in_Bank_Holding_Companies’_Decisions_37页_1mb

Summary of "The Role of Bank Capital in Bank Holding Companies' Decisions"

Core Content

This paper investigates the role of bank capital in the decision-making of bank holding companies (BHCs) in the United States, focusing on how capital constraints influence loan supply and pricing. It builds on the call option approach to bank capital introduced by Chami and Cosimano (2001), which views bank capital as a real option that allows banks to hedge against future capital constraints. The study provides empirical evidence supporting this theoretical framework and explores the implications of capital requirements on lending behavior.

Main Views and Key Findings

• Capital as a Call Option: Bank capital is treated as a real call option, where the value of the option depends on the strike price (the level of lending at which capital becomes constrained) and the volatility of loan demand. A higher capital buffer increases the strike price, reducing the value of the option, which in turn affects the loan rate.

• Optimal Capital Choice: BHCs choose their capital levels to optimize flexibility in the face of uncertainty about future loan demand. The paper uses a two-stage least squares (2SLS) approach to estimate the optimal capital ratio and its effect on loan rates.

• Empirical Evidence: The study finds that:

  • A higher optimal capital ratio leads to higher loan rates.
  • Higher loan rates result in lower loan quantities.
  • Therefore, increasing capital requirements is likely to lead to higher loan rates and a significant reduction in lending.

• Impact of Financial Crisis: After the 2007-2009 financial crisis, capital ratios showed substantial persistence, indicating that banks were more cautious in their capital management. The capital buffer increased, and the interest rate semi-elasticity of loan demand became significantly negative, meaning that a 10 basis point increase in loan rates leads to a 5.3% decrease in loan demand.

• Convexity in Capital Response: The paper argues that the real option approach implies convexity in the response of bank capital to marginal cost and revenue, which is supported by the empirical results.

• Endogeneity and Instrumentation: To address the issue of simultaneity between capital decisions, loan rates, and lending, the authors use instruments such as the target capital ratio, deposit rates, and marginal cost of lending in their regressions.

Key Information

Capital Ratios and Regulatory Requirements

• The regulatory capital requirements for U.S. BHCs are detailed in Table 1.
• The 250 largest BHCs account for 95% of total bank assets and over 90% of total bank lending as of 2013.
• These BHCs hold significant capital buffers above the regulatory minimum, with fewer than 5% of observations indicating capital constraints.

Empirical Model

• The empirical model for the capital-to-asset ratio is given by:
  $$
  \frac{K_t}{A_t} = a_0 + \left(-a_1 + a_2 \frac{K_{t-1}}{A_{t-1}}\right) \Delta \frac{K_{t-1}}{A_{t-1}} + \left(-a_3 + a_4 \frac{K_{t-1}}{A_{t-1}}\right) r_{t-1}^D + \left(-a_5 + a_6 \frac{K_{t-1}}{A_{t-1}}\right) c_{t-1}^L + \left(a_7 + a_8 \frac{K_{t-1}}{A_{t-1}}\right) Y_{t-1} + a_9 \cdot X_{t-1} + v_t^K
  $$
• The empirical model for the loan rate is:
  $$
  r_t^L = b_0 + b_1 \frac{\left(\frac{K_t}{A_t} - \widehat{\frac{K_t}{A_t}}\right)}{\widehat{\frac{K_t}{A_t}}} + b_2 r_t^D + b_3 c_t^L + b_4 Y_t + b_5 X_t + v_t^r
  $$
• The empirical model for real loan demand is:
  $$
  \ln\left(\frac{L_t}{P_t}\right) = l_0 + l_1 \widehat{r_{t-1}^L} + l_2 Y_{t-1} + l_3 X_{t-1} + v_t^L
  $$

Results

• Capital ratios are negatively related to the marginal cost of loans (including non-interest expenses and the share of nonperforming loans).
• The interest rate semi-elasticity of loan demand is zero before the crisis, but significantly negative after the crisis, indicating that higher loan rates reduce lending.
• A 1% increase in capital relative to its trend leads to a 0.25 percentage point increase in annual loan growth, suggesting that capital shortages have a negative impact on lending.
• A 23-27 basis point increase in loan rates is expected if a bank holds twice the target capital.
• A 10 basis point increase in loan rates leads to a 5.3% decrease in loan demand, consistent with findings from the U.K. by Aiyar et al. (2014).

Robustness Checks

• The paper conducts panel vector autoregression (PVAR) analysis to check the robustness of the findings.
• A one-standard deviation increase in the total risk-based capital ratio leads to an 8-12 basis point increase in loan rates within the first two quarters, resulting in a 10% drop in lending in the first quarter and a 5% drop in the second quarter.
• The results are consistent across different sub-periods, including the pre-crisis, crisis, and post-crisis periods.

Conclusion

The paper concludes that bank capital plays a significant role in the decision-making of BHCs, acting as a real call option that allows banks to hedge against future capital constraints. The empirical results support the Chami-Cosimano model, showing that capital requirements can have real effects on loan rates and lending behavior. The convexity in the capital response to marginal cost and revenue is also supported by the data. The key implication is that increasing capital requirements may lead to higher loan rates and reduced lending, which could have social costs if the cost of credit exceeds the social benefit of reduced risk-taking.