<p>This study investigates the empirical associations and predictive modeling of Risk-Based Capital (RBC) for 33 Indonesian life insurance companies over 2018 to 2024, using quarterly panel data regression and Gaussian Process Boosting (GPBoost). The Random Effects Model identifies claim expenses, firm size, and the central bank policy rate as negatively associated with RBC, while the liquidity ratio shows a consistently positive association. Premium growth and GDP growth show no significant contemporaneous effect. Return on investment is positively associated but sensitive to outlier treatment. A COVID-19 period control is included and found to exert a modest and marginally significant downward association with RBC in the primary specification, consistent with pandemic-period capital pressure among stable insurers. GPBoost, incorporating lagged RBC to capture capital persistence, substantially outperforms the panel model in forecasting future quarterly solvency ratios, reducing prediction error by approximately 43% on a common held-out test set. However, both frameworks show limited cross-firm generalization, confirming that RBC predictability is primarily a within-firm temporal phenomenon driven by each insurer’s idiosyncratic capital trajectory. The findings support the design of firm-specific, early-warning solvency monitoring frameworks and are most directly applicable to established, continuously reporting insurers rather than distressed or new-entrant firms.</p>