This paper investigates the causal relation between the performance of several cryptocurrencies and investor attention by employing Granger Causality tests and Vector Autoregression (VAR) models to examine such a relation. The findings show that interaction effects exist between returns and attention when I use Granger Causality tests, but when using VAR models, past cryptocurrency returns present a significant effect on future attention and weak reverse results. Thus, I hypothesize that if a cryptocurrency has higher past performance, then investors may pay more attention to it. After controlling the effect of economy-wide variables, I still find that past cryptocurrency returns significantly impact future attention.

Keywords: Investor attention Cryptocurrency returns Google search probability Vector autoregression