The purpose of this paper is to explore machine learning techniques to forecast the oil price. In the era of big data, we investigate whether new automated tools can improve over traditional approaches in terms of forecast accuracy. Oil price point and density forecasts are built from 22 methods, including regression trees (random forest, quantile regression forest, xgboost), regularization procedures (elastic net, lasso, ridge), standard econometric models and forecast combinations, besides the structural factor model of Schwartz and Smith (2000). The database contains 315 macroeconomic and …nancial variables, used to build high-dimensional models. To evaluate the predictive power of each method, an extensive pseudo out-of-sample forecasting exercise is built, in monthly and quarterly frequencies, with horizons from one month up to …ve years. Overall, the results indicate a good performance of the machine learning methods in the short run.Up to six months, the lasso-based models, oil future prices, and the Schwartz-Smith model provide the best forecasts. At longer horizons, forecast combinations also become relevant. In several cases, the accuracy gains in respect to the random walk forecast are statistically signi…cant and reach two-digit …gures, in percentage terms, using the R 2 out-of-sample statistic; an expressive achievement compared to the previous literature.

In this paper, a derived-demand approach is proposed to explain the positive correlation and the synchronicity between the growth rates of commodity prices and of economic activity at the global level. The focus is on important traded commodities, whose supply function is very price inelastic in the short run, such as oil and major metal commodities. The paper contributions are as follows. First, the synchronicity of oil-price and global activity cycles is presented using the tools of the common-feature literature. Second, it is shown how to improve forecasts of global activity using commodity prices, noting that one observes the latter at an almost continuous-time basis, but the former at a much lower frequency and with considerable delay. Third, the usefulness of optimal forecast combinations for oil prices is discussed employing a wide array of macroeconomic and …nancial variables. The out-of-sample R 2 statistic for model combinations can reach up to about 14%, a major improvement over the previous literature.

a b s t r a c tUsing a sequence of VAR-based nested multivariate models, we discuss the different layers of restrictions that are imposed on the VAR in levels by present-value models (PVM hereafter) for series that are subject to present-value restrictions. Our focus is novel: we are interested in the short-run restrictions entailed by PVMs (Vahid & Engle, 1993, 1997 and their implications for forecasting.Using a well-known database, maintained by Robert Shiller, we implement a forecasting competition that imposes different layers of PVM restrictions. Our exhaustive investigation of several different multivariate models reveals that better forecasts can be achieved when restrictions are applied to the unrestricted VAR. Moreover, imposing short-run restrictions produces forecast winners 70% of the time for the target variables of PVMs and 63.33% of the time when all variables in the system are considered.