Forecasting oil prices: High-frequency financial data are indeed useful

Authors: Degiannakis, S. and Filis, G.

Journal: Energy economics

Publisher: Elsevier

ISSN: 0140-9883

This data was imported from Scopus:

Authors: Degiannakis, S. and Filis, G.

Journal: Energy Economics

Volume: 76

Pages: 388-402

ISSN: 0140-9883

DOI: 10.1016/j.eneco.2018.10.026

© 2018 Elsevier B.V. The paper examines the importance of combining high frequency financial information, along with the oil market fundamentals, in order to gain incremental forecasting accuracy for oil prices. Inspired by French et al. (1987) and Bollerslev et al. (1988), who maintain that future asset returns are also influenced by past volatility, we use daily volatilities and returns from financial and commodity markets to generate real out-of-sample forecasts for the monthly oil futures prices. Our results convincingly show that although the oil market fundamentals are useful for long-run forecasting horizons, the combination of the latter with high-frequency financial data significantly improve oil price forecasts, by reducing the RMSE of the no-change forecast by approximately 68%. Results are even more impressive during the oil price collapse period of 2014–15. These findings suggest that we cannot ignore the information extracted from the financial markets when forecasting oil prices. Our results are both statistically and economically significant, as suggested by several robustness tests.

The data on this page was last updated at 19:59 on June 9, 2020.