This paper aims to investigate the predictability of technical indicators to directly forecast oil prices and compare their performances with macroeconomic variables. We find that technical indicators do exhibit statistically and economically significant in-sample and out-of-sample forecasting power under OLS regressions and forecast combinations, clearly exceeding that of well-known macroeconomic variables and state-of-the-art oil-macro forecasting variables. Moreover, the strength of the predictive evidence is substantial during recessions and expansions and can detect the typical decline in the oil returns near business-cycle peaks effectively. Furthermore, technical indicators reveal substantial economic value for investors, in terms of superior oil risk premium forecasts and sizable utility gains. The technical indicators' ability to predict the oil price stems in part from its ability to predict changes in sentiment, suggesting the financialization of oil markets.

Additional Metadata
Keywords Oil price predictability, Technical indicators, Macroeconomic variables, Out-of-sample forecasts, Business cycle
JEL Energy Forecasting (jel Q47), Energy and the Macroeconomy (jel Q43), Forecasting and Other Model Applications (jel C53)
Persistent URL hdl.handle.net/1765/105246
Journal Energy Economics
Citation
Yin, L, & Yang, Q, 524656. (2017). Predicting the Oil Prices: Do Technical Indicators Help?. Energy Economics, 56(March 2016), 338–350. Retrieved from http://hdl.handle.net/1765/105246