Bank Stability Analysis (2005–2024)

A panel-data framework for assessing financial stability of Indian Scheduled Commercial Banks using Z-score, NPAs, liquidity measures, and key risk indicators.

Overview

This repository implements a comprehensive stability analysis pipeline for Indian banks (2005–2024). It covers data collection from RBI and annual reports, metric construction (Z-score, NPAs, CAR, LDR, ALM Gap, income diversification), fixed-effects panel regression, and visual trend analysis.

Key Features

• Data Preparation: Scripts to compile, clean, and merge RBI STRBI and bank annual-report datasets.
• Metric Construction: Functions to compute Z-score, Net & Gross NPA ratios, Capital Adequacy Ratio, Loan-to-Deposit Ratio, ALM Gap, ROA, and Non-Interest Income Ratio.
• Econometric Modeling: Fixed-effects panel regression notebooks quantifying impacts of asset quality, capitalization, funding structure, and income diversification on stability.
• Visualization & Diagnostics: Notebooks for trend plots, cross-bank comparisons (public/private/foreign), and tests for heteroskedasticity and autocorrelation.

Data Description

• Source: RBI Statistical Tables Relating to Banks in India (STRBI), RBI DBIE, and bank annual reports.
• Coverage: Scheduled Commercial Banks’ annual balance sheets and ratios (2005–2024).
• Variables: Net Profit, Total Assets, Equity, Net & Gross NPAs, CAR, NIM, ROA, LDR, ALM Gap, Non-Interest Income.

Data Preprocessing

1. Compile: Aggregate key fields from RBI Excel tables and PDF reports.
2. Clean: Handle missing values, standardize definitions, adjust monetary units to constant rupees.
3. Merge: Create unified panel dataset indexed by bank and financial year.

Modeling Approaches & Hyperparameters

• Dependent Variable: Z-score = ln((ROA + Equity/Total Assets) / SD(ROA)).
• Independent Variables: Net NPA Ratio, CAR, LDR, ALM Gap, Non-Interest Income Ratio.
• Model: Fixed-effects panel regression controlling for bank-specific time-invariant effects.

Evaluation Metrics

• Regression Coefficients: Significance, sign, and magnitude of explanatory variables.
• Diagnostics: Heteroskedasticity (Breusch-Pagan), autocorrelation (Durbin-Watson), and goodness-of-fit (R²).

Execution

git clone https://github.com/your-org/bank-stability-analysis.git
cd bank-stability-analysis
pip install -r requirements.txt
jupyter notebook 01_data_prep.ipynb 02_metrics.ipynb 03_panel_model.ipynb 04_visualization.ipynb

Results & Conclusion

• Identified key drivers of bank stability: capital buffers, asset quality, funding structure, and income diversification.
• Documented temporal trends and cross-bank variations in resilience.
• Insights support macroprudential policy and bank risk-management strategies.