A1525
Title: 'One out of many': Consolidating a long-term trend forecast for investing in energy commodities
Authors: Fernanda Diaz-Rodriguez - Universidad Complutense de Madrid (Spain) [presenting]
M Dolores Robles - Universidad Complutense de Madrid (Spain)
Abstract: The aim is to add to the few studies on the annual horizon of energy price forecasting. Accurately predicting trends in energy commodity prices is vital for economic and financial stability. However, the trend itself is difficult to capture, making single forecasts unreliable. A novel approach is proposed by combining multiple trend forecasts for crude oil prices, and its effectiveness is assessed in long-term investment strategies. Combining forecasts is known to improve accuracy and reduce risk compared to relying on a single method. This approach is particularly valuable for unobservable variables like trends. Five individual trend estimation methods and seven forecast combination methods are compared. The results show that combining forecasts with a novel method called METS performs best, minimizing errors and remaining stable across market conditions. Furthermore, using estimated trends outperforms using actual prices for long-term forecasting, highlighting the limitations of short-term price predictions.
