Revealing Social Media Traces and Signals of Bitcoin

Proceedings of The 4th International Conference on Applied Research in Business, Management and Economics

Year: 2022

DOI:

[Fulltext]

Revealing Social Media Traces and Signals of Bitcoin

Dr. Tezer Yelkenci, Asst. Prof. Dr. Birce Dobrucalı Yelkenci, Prof. Dr. Gülin Vardar, Assoc. Prof. Dr. Berna Aydoğan

ABSTRACT: 

The aim of this study is to reveal the relationship between social media traces (interaction rates of Bitcoin-related tweets), social signals (content features of Bitcoin-related tweets), and Bitcoin price return. In data gathering process, Bitcoin-related tweets were collected on an hourly basis over six months. For analyzing these tweets, a two-step approach was employed. As of the first step, content features of the Bitcoin-related tweets, tweet owners’ profile features, the sentiment polarity of the tweet, and indicated emotions are assessed. In the subsequent step, for analyzing the relationship between the volatility of social media traces & social signals and the volatility of hourly Bitcoin price return, volatility spillover analysis was conducted. The empirical results designate shock transmission and volatility spillover (STVS) effects between Bitcoin and each emotion and sentiment type. The existence of bi-directional volatility spillover was found between Bitcoin and disgust, and between Bitcoin and anger. The all-remaining emotions and sentiment experienced uni-directional volatility spillover with Bitcoin, where the volatility spillovers were from the remaining emotions and sentiment to Bitcoin. According to the quartile analyses, only sentiment and anger in the first quartile, and anger in the fourth quartile indicate bi-directional volatility spillovers with Bitcoin. Nevertheless, the all-remaining variables at the all-remaining quartiles indicate a uni-directional volatility spillover running to Bitcoin. Volatility spillover from Bitcoin emerges only for fear in the fourth quartile and disgust in the first quartile. This study proposes a novel method by investigating the impact of sentiment and emotions on Bitcoin returns and volatility via two-step approach.

keywords: Bitcoin; Emotion; Multivariate GARCH; Sentiment; Volatility spillover.