Abstract

We test sentiment measured from Tweets concerning climate change and the EU emissions market, on high frequency price data. Our first main finding is that changes in the sentiment of Tweets specifically concerned with the EU emissions market predict changes in EUA prices two hours ahead with evidence of bi-directional Granger causality between changes in sentiment and changes in EUA prices. Further, we establish that periods of above (below) average sentiment correspond with periods of high (low) EUA return volatility. These findings show that sentiment does indeed have an influence in the EU emissions market. Our second finding is that while energy commodity prices, particularly NBP gas and to a lesser extent Brent oil, can account for some of the movement of contemporaneous EUA prices they are not useful at predicting these changes. This indicates that the emissions market assimilates new information from the energy market quickly. Our third main finding is that there is no evidence that Twitter sentiment concerning the general topics of climate change and global warming (rather than specifically the EU emissions market) is associated with EUA returns. This indicates that the principal means by which the EU is addressing climate change, namely the use of emissions trading, does not seem to register in the general Tweeting of Europeans about climate change and global warming