ScholarWorks@성균관대학교

초록

This paper explores the cross-border relationship between cryptocurrency prices and news sentiment by examining the Korean and global markets. We employ ChatGPT to assess sentiment in over half a million news articles from 1 December 2017 to 28 August 2023. Our findings reveal that changes in cryptocurrency prices largely drive news sentiment rather than the reverse in the global market, which differs from patterns in traditional assets and social media-driven sentiment studies. Furthermore, only Korean news sentiment predicts cryptocurrency returns in both the Korean and global markets, a predictive power that vanishes when trading volume in Korean won surpasses that in U.S. dollars. During such periods, arbitrage by Korean traders synchronizes prices between the global and Korean markets, eliminating the lead-lag relationship between sentiment and prices.

키워드