Abstract
Throughout the previous decades, the unexpected changes that commodity prices have exhibited encourage motivation for commodity price forecasting models. One of the first price forecasting models is the Black-Scholes model, and several other models are built upon this classic model. However, commodity price forecasting models nowadays lack an adequate performance when tested in long-term horizons, as seen in the current literature. This work attempts to determine a way to provide a decent accuracy in long-term (three years or more) forecasts of Brent crude oil spot price, by means of cointegration and error correction models, alongside other variables, such as other spot prices, forward prices and macroeconomic indicators. These factors have been assessed in order to ascertain which factors contribute to the forecasting approach and which of these are not helpful at all.
