Renewable energy stock prices forecast using environmental television newscasts investors’ sentiment

Abstract

The world is turning towards renewable energies to sustainably meet its increasing demand for energy. Naturally, this is being accompanied by a strong momentum in trading within the renewable energy market. Today, behavioral finance acknowledges the major role of wider psychological and social factors in shaping the stock market, through influencing investors' sentiment. Therefore, this paper explores the understudied question of whether environmental television newscasts can be used as a proxy for measuring investors' sentiment and in helping to improve the forecast accuracy of renewable energy stock prices. First, we compute the sentiment scores of the environmental newscasts of CNN, BBC News, MSNBC, and Fox News. We then use machine learning to implement a baseline forecast model, as well as an augmented one which takes the newscasts’ sentiment scores as input. Using four different accuracy metrics, we find that environmental TV newscasts can improve the forecast accuracy of renewable energy stock prices in 78 % of the experiments, and decrease the Mean Absolute Error, Mean Squared Error, and Root Mean Squared Error in 83.3 % of the experiments. We also find that the sentiments of conservative news outlets, such as Fox News, can improve the forecast accuracy of renewable energy stock prices more than liberal ones. Finally, we provide some insights into potential psychological dynamics that can help us make sense of the results, such as the negativity bias theory.