Impact of news and social media sentiments on rare earth investments

Abstract

In recent years, investing in rare earth metals has increased significantly due to its importance in producing renewable energy. Accordingly, a growing body of literature investigates the connectedness between rare earth equities and other financial markets, though very little is known about the volatility dynamics of this new asset class. Our study extends such scarce literature by employing the heteroscedastic auto-regressive (HAR) model to explore the volatility pattern of rare earth ETF. Later, we extend the baseline HAR process to explore the role of news and social media sentiment in volatility forecasting. Our empirical analysis uses the sentiment data provided by the Thomson Reuters MarketPsych Indices (TRMI) to examine whether investor sentiments collected from various media sources (i.e. traditional news media and social networking sites) would have a heterogeneous impact on the volatility predictions. Overall, our results indicate that both social and news media influence the realized volatility of rare earth equities. More importantly, this effect is positive suggesting that higher media sentiment causes higher volatility. We further observe that the extended HAR models, which include different sentiment measures such as Buzz and Sentiment, tend to outperform the baseline HAR model. Our analysis also concludes that the Buzz index has more predictive contents relative to the Sentiment index.