Abstract
We examine the predictive ability of Twitter Happiness Sentiment for six major cryptocurrencies using daily data from August 7, 2015 to December 31, 2019. At first instance, our results conclude a significant nonlinear relationship between Twitter Happiness Sentiment and cryptocurrencies. The nonlinear dependence structure is further enhanced when using the quantile-on-quantile (QQ) analysis, which indicates that high and low sentiment predicts returns of five cryptocurrencies. These findings are statistically and economically significant.
| Original language | English |
|---|---|
| Pages (from-to) | 1529-1538 |
| Number of pages | 10 |
| Journal | International Review of Finance |
| Volume | 21 |
| Issue number | 4 |
| DOIs | |
| Publication status | Published - Dec 2021 |
| Externally published | Yes |
Keywords
- cryptocurrencies
- linear and nonlinear Granger causality
- quantile-on-quantile
- Twitter Happiness Sentiment
ASJC Scopus subject areas
- Finance
- Economics and Econometrics