The relationship between bitcoin trading volume, volatility, and returns: A study of four seasons

Abstract

We study the relationship between Bitcoin trading volume, volatility, and returns using financial data for the period July 2010–November 2017. When we compare the raw annualized volatility of the Bitcoin exchange rate against common currencies, we observe that Bitcoin’s is higher. However, when the volume of Bitcoin transactions is considered, the volatility of the Bitcoin stabilizes significantly. Then we divide our sample into four distinct time periods, defined by three important events, namely, the loss of public confidence in the banking system in 2013, the MtGox Bitcoin Exchange hack in early 2014, and the introduction of the Bitcoin legislation in Japan in April 2017. Using asymmetric EGARCH models with the lag of the natural logarithm of the volume of the Bitcoin both as a regressor in the mean equation as well as in the specification of the conditional variance as multiplicative heteroskedasticity we show that volume and volatility are related after 2013, and volume and returns are related before the MtGox hack, positively and significantly. Further, during the euphoric period between the beginning of 2013 and up to the MtGox hack an unexpected rise in Bitcoin returns increases Bitcoin volatility more than an unexpected, equally sized decrease (asymmetry).