A non-linear approach to measure the dependencies between Bitcoin and other commodity markets

In this paper, we explore the relationship across cryptocurrencies and a set of commodities by using a Markov-Switching-VAR model. The parametric form of the model allows us to compute the regime-dependent impulse response functions during high and low volatility episodes and then to quantify bidirectional spillovers between both markets. Our main results show that responses to commodity shocks are more important in the high volatility regime for almost all commodities. However, we find a very moderate impact of the Bitcoin fluctuations on commodities, although situations seem to differ according to the commodity.