Spillover shifts in the FX market: Implication for the behavior of a safe haven currency

This paper explores the behavior of safe haven currencies by analyzing shock transmission among major currencies. To capture state-dependent directional spillovers, we incorporate Markov regime-switching parameters into the spillover model and estimate them using a Bayesian MCMC algorithm. By considering weekly data from September 2000 to March 2020, we find that the Japanese yen and the Swiss franc, both of which yield relatively high excess returns in times of crisis, exhibit larger reductions of shock transmission and reception during periods of high-volatility than during periods of low-volatility. This implies that the safe haven currencies insulate themselves from shocks from other currencies by reducing interdependence across the FX market in crisis.