Forecasting Volatility in Cryptocurrency Markets

Segnon Mawuli, Bekiros Stelios

Abstract
In this paper, we revisit the stylized facts of cryptocurrency markets and proposevarious approaches for modeling the dynamics governing the mean and varianceprocesses. We first provide the statistical properties of our proposed models and studyin detail their forecasting performance and adequacy by means of point and densityforecasts. We adopt two loss functions and the model confidence set (MSC) test toevaluate the predictive ability of the models and the likelihood ratio test to assess theiradequacy. Our results confirm that cryptocurrency markets are characterized by regimeshifting, long memory and multifractality. We find that the Markov switching multifractal(MSM) and FIGARCH models outperform other GARCH-type models in forecastingbitcoin returns volatility. Furthermore, combined forecasts improve upon forecasts fromindividual models.

Keywords
Bitcoin; Multifractal processes; GARCH processes; Model confidence set; Likelihood ratio test