Variance-Based Spillover Analysis between Stock Markets: A Time Varying Parameter Approach

Abstract

This paper proposes a variance-based spillover impact analysis embedded with a dynamic Kalman ltering in order to detect a causality relationship from the US stock markets into the European and emerging stock markets during the nancial crisis. It has mainly two new contributions to the literature. Firstly, it uses variance rather than returns to analyze the spillover impact between the markets. Secondly, and more importantly, it is an econophysics research as it examines causality relationship with the Kalman ltering in physics. We calculate time-dependent conditional stock market variances for Dow Jones, DAX, FTSE, RTS (Russia), and BIST (Turkey) by employing SWARCH model. The empirical analysis examines the causal relationship between Dow Jones into the other stock markets employing Granger causality tests in order to detect the direction of volatility spillover relationship. As an embedded analysis, we follow a dynamic approach by using the Kalman ltering as a time varying parameter model to depict the time varying interaction between stock markets volatilities. The empirical results point out unidirectional Granger causality from Dow Jones to the other markets indicating the spillover impact of the volatility starting from the US markets and expanded into the world in the latest global crisis.