We examine the dynamics of the signed-spillover across financial markets using historical decomposition approach. By incorporating Markov-switching framework into the VAR model, this paper assesses the dynamics of signed-spillover during turbulent periods and period of tranquillity. Additionally, this approach enables us to detect the source and direction of the spillover and identify its signs. We show that this approach outperforms the classical single-regime spillover estimation by distinguishing shocks under different economic conditions. Specifically, we assess spillovers in global financial markets using realised variance between January 1999 and December 2017. Our empirical findings clearly indicate that spillovers are intense during period of turbulence and moderate during periods of tranquillity.