Analyzing the Relationship between the Dow Jones Index and Oil Prices Using the ARIMAX Model

Abstract

The values prediction of the Dow Jones index is essential in the global financial markets systems. The index provides a clear vision of what is happening in the market as a whole. Hence, it offers integrated information on that index to stipulate forecasts characterized as efficient for investors and shareholders. In this study, the ARIMAX model was used to predict the daily Dow Jones index values from 1/1/2020 to 1/ 5/2020 (the spread of COVID-19), considering Brent crude's effect daily prices as an external factor. The Dow Jones Index daily price prediction process went through several stages. The first stage is the time series stationary test phase through the Augmented Dickey-Fuller test. The second stage is achieving stationary by taking the first difference, passing through the stage of identifying the model, and determining the rank based on criteria (AIC), (BIC), (RMSE). The preference of the model was shown in ARIMA (0,1,2) for the Dow Jones index series. The ARIMA (1,1,0) model was shown for crude price Brent series and determining the order of the transfer function of the ARIMAX model. The comparison stage between the models ARIMA (0,1,2) and ARIMAX(3,1,1)(0,0,1) by residuals scatter plot and (Ljung-Box) test for each model. The results demonstrated the superiority of the ARIMAX model over the ARIMA model. The daily Dow Jones Index values were predicted based on corresponding Brent crude prices according to the ARIMAX (3,1,1) (0,0,1) model. The researchers did not find substantial differences in the index’s behavior, except for a slight decrease in the index's value.