Regime switches and commonalities of the cryptocurrencies asset-class

Abstract

In this paper we test for regime changes in the price dynamics of Bitcoin, Ethereum, Litecoin and Monero, as representatives of the cryptocurrencies asset class. Data are observed daily from January, 1, 2016 to October, 15, 2019. Best specifications within Gaussian and Autoregressive Hidden Markov models for price differences are selected through the AIC and BIC information criteria by considering up to four hidden regimes. The empirical results suggest that at most three common states may be considered for the basket of cryptocurrencies under investigation; a fourth state may be relevant as an added factor to the dynamics description of the individual cryptocurrencies rather than to the whole basket. Finally, we test the out-of-sample performance of estimated regime switching models; optimal results, in terms of RMSE and correlation between predicted and real values, are obtained in the case of two common or three individual regimes.

Keywords: Cryptocurrencies, Regime Switching, Hidden Markov Models, Forecasting Analysis

JEL Classification: C22, C52, C53, C58