Forecasting MSCI World Energy Sector Index with the SARIMA Model

Abstract

Energy plays a crucial role in economic systems in terms of consumption and production. At the present, there is a high fluctuation of energy prices due to the business cycle movement and the differences in business energy consumption demanding in each cycle; therefore, the energy index prediction could help investor plans appropriately. The model used in this study is Seasonal Autoregressive Integrated Moving Average model (SARIMA). This is a model increasing seasonal effects which was developed from ARIMA (p, d, q) of Box and Jenkins. The purposes of this study are 1) to construct a suitable model for MSCI World Energy Index by using SARIMA (Seasonal Autoregressive Integrated Moving Average), and 2) to compare Forecast Accuracy of MSCI World Energy Sector Index via SARIMA Model. The data in this study is a monthly information from the MSCI World Energy Index from 2005 to 2019 (15 years). In the research methodology, there is a data stationary tested by using the unit root test, and simulating SARIMA model. After selecting the most appropriate model, the data prediction test was operated. In conclusion, the result of this study revealed that the most appropriate model for prediction was SARIMA (2,1,1)×(2,1,3)₁₂. The prediction model outcome was very close to the real indices, when the deviation of RMSE was 16.68 and MAE was 12.39 respectively.