Forecasting the Price of Crude Oil with Multiple Predictors

For the price of crude oil, this paper aims to investigate the predictive content of a variety of variables including oil futures prices, exchange rates of particular countries and stock-market indexes. Out-of-sample forecasting results suggest that oil futures prices have marginal predictive power for the price of oil at a 1-month forecast horizon. However, they generally lose their forecasting power at higher forecast horizons. The results also suggest that exchange rates help predicting oil prices at higher forecast horizons. The paper also considers forecast averaging and variable selection methods, and fınds that forecast averaging significantly improves the forecasting performances