Dynamic Price and Volatility Relationships between Crypto Returns and Twitter-based Economic Uncertainty

Abstract

In this study we examined the mean and the volatility transmissions of Twitter-based uncertainty and returns of Bitcoin, Ethereum, BNB, XRP, and Cardano in a time-varying context and considered the structural changes. The availability of the data directs us to have a different starting date for the analysis of five cryptocurrencies. However, our data span for all types of cryptocurrencies ended on August 31, 2022. We made several findings. First, our findings indicate that Twitter-based uncertainty has a more robust predictive power on cryptocurrency returns in the time-varying context compared with the standard Fourier-Granger causality results. Second, the empirical findings of the volatility transmission analysis also provide similar results. Third, the findings of the dynamic Fourier variance causality analysis exhibit robust evidence for the bidirectional volatility spillovers between the uncertainty indicator and the cryptocurrency returns, which in turn suggests a significant risk transmission.

JEL: C22, G11, G15.