Econometric models for oil price forecasting: A critical survey

In this paper we investigate the existing econometric literature on oil price forecasting. In particular, we: i) develop a taxonomy of econometric models for oil price forecasting; ii) provide a critical interpretation of the different methodologies; iii) offer a comprehensive interpretation and justification of the heterogeneous empirical findings in published oil price forecasts. We distinguish among three different categories of econometric models: time series models; financial models; structural models. We find that the type of econometric model used seems to affect the type of forecasts that is produced but, despite the relatively large number of criteria which are available to evaluate the forecasting performance of each proposed model, it is not possible to identify which class of models outperforms the others in terms of forecasting accuracy.