Time Varying Volatility Indexes and Their Determinants : Evidence from Developed and Emerging Stock Markets

This paper investigates volatility spillovers across 16 stock markets of both advanced and emerging economies using the spillover index methodology put forward by Diebold and Yilmaz (2012). Realised volatility as defined by Andersen et al (2003) calculated from high frequency data form the basis on which these spillovers are calculated. They are compared with spillovers based on the volatility estimators put forward by Garman and Klass (1980), Parkinson (1980) and the univariate GARCH methodology (Bollerslev, 1986) used in many previous studies. We find that the time series of total spillovers is similar regardless of the volatility proxy used and spillovers increased dramatically during the global financial crisis of 2008 and the European sovereign debt crisis that followed. More differences arise when comparing directional spillovers to and from individual stock markets, particularly when using GARCH based estimations. We find that the larger stock markets from the advanced western economies, particularly the US, dominate volatility transmission to other markets. Emerging markets such as China, India and Brazil are still relatively isolated and contribute less to global volatility spillovers, though their contribution has increased considerably after 2006. We investigate potential determinants of net spillovers between markets and find that the level of volatility in one market relative to that in other markets is the most important factor in increasing spillovers transmitted