Modelling Volatility Dynamics of Cryptocurrencies Using GARCH Models

Abstract

Cryptocurrencies have become increasingly popular in recent years at- tracting the attention of the media, academia, investors, speculators, regu- lators, and governments worldwide. This paper focuses on modelling the

volatility dynamics of eight most popular cryptocurrencies in terms of their market capitalization for the period starting from 7th August 2015 to 1st August 2018. In particular, we consider the following cryptocurrencies; Bitcoin, Ethereum, Litecoin, Ripple, Moreno, Dash, Stellar and NEM. The GARCH-type models assuming different distributions for the innovations

term are fitted to cryptocurrencies data and their adequacy is evaluated us- ing diagnostic tests. The selected optimal GARCH-type models are then used

to simulate out-of-sample volatility forecasts which are in turn utilized to es- timate the one-day-ahead VaR forecasts. The empirical results demonstrate

that the optimal in-sample GARCH-type specifications vary from the selected

out-of-sample VaR forecasts models for all cryptocurrencies. Whilst the em- pirical results do not guarantee a straightforward preference among

GARCH-type models, the asymmetric GARCH models with long memory property and heavy-tailed innovations distributions overall perform better for all cryptocurrencies.