Testing for time-varying stochastic volatility in Bitcoin returns

Abstract

The study will be the first to offer empirical justification for time-varying stochastic volatility in Bitcoin returns. Specifically, it tests for time variation in both the trend and transitory components of the stochastic volatility using the unobserved components model that accounts for same. Thereafter, it calculates the Bayes factor using the approach of Chan (2018) which involves the Savage-Dickey density ratio in order to avoid the computation of the marginal likelihood. The results overwhelmingly support at least one time-varying stochastic volatility component in Bitcoin returns and the transitory component is favoured in this regard. These results are robust to different data frequencies.