The risks of cryptocurrencies with long memory in volatility, non-normality and behavioural insights.

This paper aims to study the impacts of long memory in conditional volatility and conditional non-normality on market risks in Bitcoin and some other cryptocurrencies using an Autoregressive Fractionally Integrated GARCH model with non-normal innovations. Two tail-based risk metrics, namely Value at Risk (VaR) and Expected Shortfall (ES), are adopted to study the tail behaviour of market risks in Bitcoin and some other cryptocurrencies. Empirical investigations for the tail behaviour based on real exchange rate data of cryptocurrencies are conducted. An extreme-value-theory-based approach is used to study potential improvements in the estimation for the risk metrics under GARCH-type models. The possibility of explosive regimes in cryptocurrencies' volatilities is examined using Markov-switching GARCH models. [ABSTRACT FROM AUTHOR]