Fundamentally Sentimental: A Cointegration Approach to Disentangling Forces of Demand in the Bitcoin Market

Abstract

This thesis employs a vector error correction model (VECM) to investigate the dynamic relationships between forces of demand in the bitcoin market. We contribute to the literature regarding media-based sentiment and financial time series by constructing a novel, attention-weighted measure of public sentiment towards bitcoin. We situate sentiment alongside price, trading volume, network transaction volume, active addresses, and hashrate as endogenous variables in a multivariate representation of the bitcoin market. The VECM allows us to account for the presence of cointegration between variables, allowing for the interpretation of both long-run and short-run interdependencies between bitcoin prices, sentiment, and fundamentals. Our empirical results show that there is a lot more at play in bitcoin price formation than previously suggested by the literature, as sentiment, fundamentals, and trading volume are all found to be significantly associated with price at various time horizons. We also provide evidence for the presence of a feedback mechanism in the bitcoin market, as sentiment and fundamentals appear to depend on each other in the long run, while also being significantly affected by short-term fluctuations in market behavior.