ESTIMASI HARGA BATUBARA MENGGUNAKAN MODEL HYBRID ARIMA-GARCH = ESTIMATION OF COAL PRICES USING HYBRID ARIMA-GARCH MODELS

The estimation of coal prices is an important aspect in Indonesia's energy industry, considering that significant price fluctuations can have an impact on business decisions and economic policies. This research aims to estimate the best hybrid ARIMA-GARCH model and predict coal prices using the hybrid ARIMA-GARCH model for September 2024 to January 2025. The ARIMA method is used to catch trend and seasonal patterns in coal price data, while the GARCH method is used to properly model price volatility that often appears in financial data. This method combination is expected to produce more accurate predictions compared to a single method. The Hybrid ARIMA-GARCH model is a combination model of the ARIMA and GARCH models, which can be used to overcome the residual problems of ARIMA models that indicate heteroscedasticity in residual variance (volatility). The steps in this analysis and discussion are descriptive statistics, stationarity testing, forming the best ARIMA model, forming the best GARCH model, combining hybrid ARIMA-GARCH models, determining the best hybrid ARIMA-GARCH model, measuring the accuracy of hybrid ARIMA-GARCH forecasting, and forecasting. This research was performed using Eviews in analyzing reference coal price data. The outcome of this research is that the best model for coal price is hybrid ARIMA(3,1,1)-GARCH(1,1) with MAPE value = 9.76%. According with the formula of the ARIMA(3,1,1)-GARCH(1,1) model is D(HBA)=4.474 log⁡〖z_(t-1) 〗+4.973 log⁡〖z_(t-2) 〗+0.001+ 0.146α_(t-1)+ 0,0,596α_(t-2)-0.005+ 0,057ε_(t-1)^2+ 0,221σ_(t-1)^2. Based on the best model, the forecasting outcomes for September 2024 to January 2025 are $115,7; $120,83; $118,99; $129,2; $117,31, respectively, which indicates that coal prices in June to October 2024 have decreased in price.