Negativity Bias in the European Emissions Market: Evidence from High-Frequency Twitter Sentiment

Abstract

High-frequency sentiment time series are extracted from Twitter data concerning the European emissions market and are used to explain returns and volatility in European emissions futures. The measures of negative sentiment are shown to Granger-cause EUA futures returns, while positive sentiment does not, indicating the presence of a negativity effect. Using a Threshold GARCH model, it is shown that periods of strong (weak) sentiment correspond with periods of high (low) volatility. Our findings are significant for understanding the role of sentiment in European emissions markets with implications for other markets subject to uncertain political actions and public opinions.

Keywords: Emissions Market, Sentiment, Twitter

JEL Classification: G41, Q52, G13