Evaluating the empirical performance of alternative econometric models for oil price forecasting

The empirical literature is very far from any consensus about the appropriate model for oil price forecasting. Several specifications have been proposed: some concentrate on the relationship between spot and futures prices ("financial" models), while others assign a key role to economic fundamentals ("structural" models). In this work we systematically test and evaluate the ability of several alternative econometric specifications to capture the dynamics of oil prices. Moreover, we propose a new class of models which combines the relevant aspects of financial and structural specifications ("mixed" models). We evaluate the forecasting performance of each class of models using different measures of forecast accuracy. We also analyse the effects of different data frequencies on the coefficient estimates and forecasts of each selected specification. Our empirical findings suggest that financial models are to be preferred to time series models. Both financial and time series models are better than mixed and structural models. Although the random walk model is not statistically outperformed by any of the alternative models, our empirical results suggest that theoretically well-grounded financial models are valid instruments for producing accurate forecasts of the WTI spot price.