Time-Varying Volatility in Bitcoin Market and Information Flow at Minute-Level Frequency

In this article, we analyze the time series of minute price returns on Bitcoin market through statistical models generalized autoregressive conditional heteroscedasticity (GARCH) family. We combine an approach that uses historical values and their volatilities—GARCH family models, with a so-called Mixture Distribution Hypothesis, which states dynamics are governed by information flow about market. Using Bitcoin-related tweets, trade volume, bid–ask spread, as external signals, test for improvement in volatility prediction several GARCH model variants minute-level series. Statistical tests show GARCH(1,1) cGARCH(1,1) react best to addition signals process out-of-sample data.