The impact of news media on Bitcoin prices: modelling data driven discourses in the crypto-economy with natural language processing

Abstract

This paper examines the relationship between events reported in international news via categorical discourses and Bitcoin price. Natural language processing was adopted in this study to model data-driven discourses in the crypto-economy, specifically the Bitcoin market. Using topic modelling, namely Latent Dirichlet Allocation, a text analysis of cryptocurrency articles ( N = 4218) published from 60 countries in international news media identified key topics associated with cryptocurrency in the international news media from 2018 to 2020. This study provides empirical evidence that across the corpora of international news articles, 18 key topics were framed around the following categorical macro discourses: crypto-related crime, financial governance, and economy and markets. Analysis shows that the identified discourses may have had a ‘social signal’ effect on movements in the crypto-financial markets, particularly on Bitcoin's price volatility. Results show these specific discourses proved to have a negative effect on Bitcoin's market price, within 24 h of when the crypto news articles were published. Further, the study found that in some cases, the source of the news may have amplified the volatility effect, particularly in terms of geographical region, relative to broader market conditions.