Bank Capital Regulation: A Dynamic Analysis

The study's core is a dynamic model investigating how bank behavior changes with increased government surveillance. The model's economy consists of separate 'islands' (representing industries or regions), each with a bank holding a credit monopoly. Crucially, each island faces its own unique, undiversifiable profitability shock. Banks in the model face idiosyncratic profitability shocks to their assets, and equity issuance is costly due to information asymmetry. Equity capital absorbs losses, protecting charter value during negative asset shocks. Unlike most previous studies that model capital regulation as always binding, this model considers it 'occasionally binding,' allowing for capital buffer accumulation through retained profits, irrespective of enforcement. Stricter regulatory surveillance prompts banks to increase these buffers to avoid violating capital requirements. This 'occasionally binding' approach allows for a more nuanced analysis of bank behavior under varying levels of regulatory pressure and enforcement.

A key distinction in the model lies in contrasting incomplete and complete enforcement of capital regulations. The model acknowledges that before the introduction of Prompt Corrective Action (PCA) programs, capital regulations were often 'softly implemented' or 'incompletely enforced' – regulators sometimes overlooked undercapitalized banks. However, PCA programs increased regulatory scrutiny, requiring rigorous self-assessment of assets. The model captures this shift by varying the 'inspection probability' – the likelihood of regulatory review. This probability acts as a proxy for the strength of regulatory pressure. An increase in this probability from incomplete to complete enforcement is the focus of the dynamic analysis. The model's novelty lies in explicitly modeling the strength of regulatory pressure as an inspection effort within the banking system dynamics.

The model predicts that strengthening regulatory surveillance, modeled as an increase in the inspection probability, will have different short-run and long-run effects on banks. In the short run, the model predicts a decrease in banks' charter value (market value of capital) and lending, particularly impacting highly leveraged and profitable banks. These banks experience a larger reduction in value because the capital requirements become more binding for them. However, the long-run effect is a stabilization of the banking system as banks accumulate sufficient regulatory capital to no longer be capital-constrained, lowering the overall default rate. The model captures the short-run cost (decreased lending and charter value) and the long-run benefit (increased system stability) of stricter capital regulation. This provides a theoretical foundation for understanding the complex relationship between regulatory pressure and bank behavior.

The dynamic analysis simulates the transition from incomplete to complete enforcement of capital regulations—an increase in the inspection probability from 0.26 to 1. This analysis tracks aggregate variables like loan totals, default rates, and capital ratios over time. The model reveals that initially, strengthening surveillance causes a capital crunch and financial instability due to decreased charter value and increased default incentives. However, over the long run, the gradual increase in regulatory capital leads to a more stable banking system with a lower default rate than observed under incomplete enforcement. The model demonstrates different short-run versus long-run implications of regulatory tightening. While the short run shows a decrease in lending and charter value, the long run indicates a stabilized banking system due to higher regulatory capital. This dynamic analysis is crucial for assessing the long-term effectiveness of regulatory reforms.

This section details the use of Japan's Prompt Corrective Action (PCA) program as a natural experiment to test the theoretical model's predictions. The PCA, implemented in stages (preliminary in FY 1997, full effect in April 1998), introduced a self-assessment process for banks, holding them accountable for realistic asset valuations. This process, subject to external audit and regulatory review, aimed to lessen forbearance and strengthen existing regulatory frameworks. The PCA's introduction is considered a significant strengthening of bank capital supervision. The study leverages this policy change to analyze its causal impacts on bank behavior, using the pre-PCA period (before FY1997) as a control group and the post-PCA period (FY1997 onwards) to assess the impact of heightened regulatory scrutiny. The creation of the Financial Services Agency (FSA) in April 1998, with its increased independence and credibility compared to the Ministry of Finance, further reinforced the PCA's impact.

The empirical analysis uses various specifications to examine the impact of the PCA on several bank outcome variables: regulatory capital adequacy ratios, regulatory capital buffers (difference between reported and target ratios), market capital ratios (proxy for franchise value), and loan growth. The data comes from two main sources: Nikkei Digital Media Inc. Bank-level panel data covers FY 1996-1997, consisting of financial statements from 116 Japanese banks listed on Japanese stock exchanges (total sample size of 232). Matched bank-firm loan data from the same source allows for analysis at the loan level, encompassing approximately 2500 firms per year. The loan data includes both short-term and long-term loans from various Japanese banks, including city, trust, regional, and mutual banks. The authors discuss challenges in handling loan-level data due to bank mergers and restructurings, and describe methods used to address these issues, ensuring accurate tracking of loan amounts across time periods.

The study employs specific estimation methods to analyze the data. Bank-level data is analyzed using a conventional within estimation method with two-way fixed effects for the FY 1996 and 1997 periods. The analysis incorporates one-period-lagged values of regulatory capital buffers, return on assets (ROA), and control variables (size, overseas branches, Tobin's Q). For the loan-level analysis using matched lender-borrower data, a three-way fixed effects model is employed. This analysis addresses survivorship bias—the non-random continuation of bank-firm lending relationships—using Heckman's two-stage regression technique. The first stage involves a probit regression to model the probability of relationship continuation, and the second stage incorporates the inverse Mills ratio from the probit model into the loan-level regression. To handle the potential for time-varying coefficients, the probit model is estimated year by year (FY 1996 and 1997). The authors detail the inclusion of various control variables to ensure robust causal inference.

The empirical findings strongly support the theoretical predictions. The analysis reveals that the preliminary implementation of the PCA in FY 1997 significantly increased regulatory capital, particularly for highly leveraged banks. Concurrently, it significantly decreased the market capital ratios (franchise value), again most prominently for highly leveraged and profitable banks. This aligns with the model's prediction of short-run decreased charter value. Placebo tests, using data from FY 1993-1996, confirm the robustness of these findings. The impact of the PCA was only statistically significant in FY 1997, indicating the PCA's specific influence on bank behavior. The results demonstrate that the increased regulatory pressure from the PCA program led to a short-term contraction in lending and a decrease in banks' market value, alongside an increase in regulatory capital. This supports the study's hypothesis about the short-run consequences of strengthening capital supervision.

The empirical analysis focuses on the short-run implications of strengthened bank capital supervision, using Japan's Prompt Corrective Action (PCA) program as a natural experiment. The PCA, initially implemented in FY 1997, aimed to reduce regulatory forbearance by mandating rigorous self-assessment of bank assets. The study uses bank-level and loan-level data from FY 1996 and 1997 to analyze the immediate effects of this regulatory change. The results indicate that the PCA led to a decrease in lending and charter value (market value of capital), especially for highly leveraged and profitable banks. This finding aligns with the short-run predictions of the theoretical model which showed that stricter regulatory oversight leads to a reduction in both lending and franchise value, especially for those banks that were highly leveraged. This short-term contraction, however, coexists with a simultaneous increase in regulatory capital, indicating that banks adjusted their behavior by increasing their regulatory capital buffers, primarily by retaining earnings. The decrease in franchise value supports the observation by Sarin and Summers (2016) regarding the short-term costs associated with stricter capital regulation.

While the short-run impact of strengthened bank capital regulation involves decreased lending and franchise value, the study suggests a positive long-run consequence: increased financial stability. The dynamic model predicts that the accumulation of regulatory capital resulting from stricter supervision will ultimately reduce the risk of bank defaults. Though not directly tested empirically in this specific section, this long-run effect is supported by the theoretical model and contrasted with the observed short-run effects. The model demonstrates that while the short-run response may involve a credit crunch and decreased bank value, the long-run outcome is a lower default rate and a more stable banking system. This suggests that the initial costs of stricter capital requirements are potentially offset by long-term improvements in systemic stability. The empirical data supports this by showing that the increased regulatory capital observed after the PCA's implementation might translate into a lower long-run default rate for Japanese banks.

The study concludes by synthesizing the short-run and long-run implications of strengthened bank capital regulation. The empirical findings confirm the short-run observation by Sarin and Summers (2016) that stricter capital regulations, while increasing regulatory capital, can simultaneously decrease bank franchise value, potentially causing short-term financial instability. However, the authors emphasize that the long-run perspective reveals a counterbalancing effect. The model's dynamic analysis shows that after the initial short-run adjustments (contraction in lending and decreased franchise value), the banking system will stabilize in the long run as banks accumulate regulatory capital and reduce default risks. This dynamic view is crucial for policymakers, highlighting the importance of considering both short-term costs and long-term benefits before implementing strict capital regulations. The Japanese PCA program serves as a real-world example of this complex interplay between short-run adjustments and long-run stability.