Title: Networks of volatility spillovers among stock markets
Authors: Eduard Baumohl - University of Economics in Bratislava (Slovakia) [presenting]
Stefan Lyocsa - University of Economics in Bratislava (Slovakia)
Tomas Vyrost - University of Economics in Bratislava (Slovakia)
Evzen Kocenda - Charles University (Czech Republic)
Abstract: In a network analysis of 40 developed, emerging and frontier stock markets during the 2006-2014 period, volatility spillovers during both the global financial crisis and tranquil periods are described and modeled. The resulting market interconnectedness is depicted by fitting a spatial model incorporating several exogenous characteristics. We document the presence of significant temporal proximity effects between markets and somewhat weaker temporal effects with regard to the US equity market - volatility spillovers decrease when markets are characterized by greater temporal proximity. Volatility spillovers also present a high degree of interconnectedness, which is measured by high spatial autocorrelation. This finding is confirmed by spatial regression models showing that indirect effects are much stronger than direct effects; i.e., market-related changes in "neighboring" markets (within a network) affect volatility spillovers more than changes in the given market alone, suggesting that spatial effects simply cannot be ignored when modeling stock market relationships. Our results also link spillovers of escalating magnitude with increasing market size, market liquidity and economic openness.