Fundamental Sentiment and Cryptocurrency Risk Premia

Abstract

This paper investigates the cross-sectional predictive ability of text-based factors in the cryptocurrency market, an important asset class for retail and institutional investors. We employ Bidirectional Encoder Representations from Transformers (BERT) topic modeling to analyze news articles discussing the top 43 cryptocurrencies by market capitalization. We build text-based factors related to cryptocurrency fundamentals and find that the exposure to sentiment on fundamentals is priced. Cryptocurrencies that have a high beta with respect to our measure of fundamental sentiment are typically platform and general payment tokens, whereas low beta currencies are governance tokens. Currency betas correlate with measures of value on the blockchain, such as the ratio of the network of users and transactions to market cap. High beta currencies are more sensitive to fundamental news and earn a risk premium. Our results demonstrate the importance of considering text-based factors when analyzing cryptocurrency returns.

Keywords: cryptocurrency, fundamentals, media coverage, textual analysis

JEL Classification: G11, G12, G14, G32