Modeling the Gross Returns of the Big Four Banks’ Stocks in Indonesia Using a Generalized Linear Model (GLM) with the AIC, BIC, and Deviance Approaches

Abstract

This study examines the macrofinancial sensitivities and returns dynamics of the four systemically important banks in Indonesia using a Generalized Linear Model framework optimized via operational parameter criteria. Utilizing daily returns data, the empirical framework evaluates the transmission channels of domestic equity index movements, foreign exchange fluctuations, and central bank monetary policies while explicitly addressing non-Gaussian error distributions. The statistical estimations demonstrate that domestic capital market performance acts as the definitive anchor for banking stock returns, although public and private banking instruments exhibit divergent elasticity profiles reflecting distinct institutional mandates and funding structures. Furthermore, residual diagnostics reveal a complete rejection of error normality across all asset vectors, confirming that traditional linear estimation models produce inefficient standard errors under conditions of extreme market volatility. By accommodating alternative exponential distribution profiles and flexible non-linear link functions, the generalized framework minimizes unexplained variance and yields highly consistent parameter metrics. These findings underscore the critical utility of non-linear structural modeling in accurately identifying systematic risks, optimizing corporate banking portfolios, and strengthening macroprudential surveillance systems within emerging financial architectures.