Gold is one of precious metals which had been used widely as a kind of currency till August 15, 1971. The gold price has been steadily soaring since mid 2004 from USD 391.40 in July 2004 to USD 1813.50, a historically high price in August 2011. There are certainly a couple of reasons for its hike.
This study uses ARIMA, VAR and exponential smoothing models to forecast the trend of gold price. Oil price, S&P 500 index and USD index are also used in the multivariate VAR model. The data are collected at the monthly interval covering the span from January 1995 to March 2014. The data from January 1995 to December 2012 are used for estimating the models while the rest of the data are used as forecasts.
First, the four variables are found integrated of order one by using ADF and PP unit root tests. The ARIMA, VAR and exponential smoothing models are next estimated from in-the-sample data, and then are used for forecasts. Finally, four indexes are used to select the best forecast model. They are root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE) and Theil inequality coefficient. The evaluation result shows that exponential smoothing model has the best forecasting ability of all.
