Official media sentiments toward energy and equity returns: Evidence from China

In this study, we constructs a set of novel sentiment indices derived from energy-related articles published by leading official media outlets in China using multiple machine learning techniques. Our results show that these official media sentiment indices consistently and significantly predict returns of both the China Securities Index (CSI) Energy Index and the Energy Exchange-Traded Fund (ETF). Furthermore, we find that official media sentiment conveys fundamental energy-related information not yet reflected in asset prices. Based on these return patterns, we design several trading strategies and show that they generate substantial economic value for both equity-focused and mean-variance investors. Overall, the findings underscore the influential role of official media as a channel for government and policymakers voices in shaping equity returns in China's energy sector. • Sentiment indices constructed from Chinese official energy-related news. • Sentiment measured using multiple machine learning techniques. • Indices predict returns of China's energy stock index and exchange-traded funds. • Official media sentiment conveys fundamental energy information early. • Sentiment-based trading strategies deliver notable economic value.