Volatility and return jumps in bitcoin

Abstract

In this article we are interested in understanding the dynamics of Bitcoin daily returns and volatility. Cryptocurrencies have very high unconditional volatility, and are subject to sudden, massive, price swings. We start with a standard log-normal stochastic volatility model, then explore two formulations which incorporate discontinuous jumps to volatility and returns. Jumps to volatility are permanent, while jumps to mean returns have contemporaneous effects only. Results point to two high volatility periods: the first from late 2013 to early 2014, likely associated to the Mt. Gox incident; the second covers the year of 2017, peaking on December — likely driven by increased popular attention. Jumps to mean returns are specially relevant to capture large price variations, mostly negative, associated with formative events in cryptocurrency markets, such as hacks and unsuccessful fork attempts.