Time Series Innovation: Leveraging BetaSutte Models to Enhance Indonesia's Export Price Forecasting

Abstract

This study introduces a novel application of the Modified Trend-Augmented α-Sutte Indicator (BetaSutte) model for forecasting Indonesia's export prices and compares its performance with the traditional ARIMA approach. Accurate export price forecasting is crucial for economic planning, trade policy formulation, and business strategy development in Indonesia's dynamic and globally connected economy. Using monthly export value data from January 2022 to September 2024 obtained from Indonesia's Central Bureau of Statistics (BPS), we examined whether the BetaSutte model's decomposition of trend and residual components offers enhanced predictive accuracy over the conventional ARIMA methodology. Results show that while the ARIMA(0,1,0) model demonstrated superior in-sample performance (Training MAPE: 7.71% vs. 80.78%), the BetaSutte model achieved better out-of-sample forecasting accuracy (Testing MAPE: 11.22% vs. 11.61%). The BetaSutte model's linear trend component identified a negative slope (coefficient: -158.4), indicating a systematic decline in Indonesia's export values over the study period, which has important implications for trade policy. Furthermore, the model successfully captured the volatility in export prices through its residual forecasting component. These findings suggest that the BetaSutte model's explicit modeling of trend components provides meaningful advantages for export price forecasting, despite its more complex implementation. This research contributes to the growing literature on hybrid forecasting methodologies and offers practical guidance for stakeholders interested in Indonesia's international trade dynamics. For policymakers, the results highlight potential challenges for Indonesia's export competitiveness and suggest the need for targeted interventions to address the identified downward trend in export values.