Ansari Saleh Ahmar, Agung Triutomo, Nur Ikhwana
This study develops a hybrid decision support framework that integrates Triple Exponential Smoothing with GARCH models to enhance inflation forecasting accuracy for monetary policy management in Indonesia. Using headline Consumer Price Index (CPI) monthly inflation data from January 2010 to September 2023, validated through out-of-sample rolling window testing with 80/20 train-test split, the study address the critical need for robust forecasting methodologies in economic policy decision-making. The hybrid approach leverages Triple Exponential Smoothing's capability to capture trend and seasonal patterns while incorporating GARCH models to account for volatility clustering in inflation data. Results demonstrate that the hybrid model outperforms standalone Triple Exponential Smoothing and GARCH(1,1) models with a 27.8 % reduction in Mean Absolute Percentage Error (MAPE) and enhanced variance prediction. The hybrid framework integrates TES for trend-seasonal decomposition with GARCH modeling of residual volatility, creating synergistic benefits beyond simple model averaging. The 95 % Value-at-Risk estimates from the hybrid model achieve 94.5 % coverage rate with narrower prediction intervals than conventional methods. Implementation of this framework enables policymakers to identify emerging inflation risks 2–3 months earlier than traditional methods, determined through systematic back-testing analysis comparing signal timing across inflation threshold events (>1.5 % point increases) than traditional methods, thereby expanding the window for calibrated policy responses. This research contributes to the decision sciences literature by establishing an integrated approach that bridges statistical forecasting techniques with practical monetary policy management applications. © 2026 Elsevier B.V.
Department of Statistics, Faculty of Mathematics and Natural Sciences, Universitas Negeri Makassar, Makassar, Indonesia
