The Nature of Volatility Spillovers Across the International Capital Markets

Abstract

This paper studies the nature of volatility spillovers across countries from the perspective of network theory and by relying on data of US-listed ETFs. I use a Lasso-related technique to estimate the International Volatility Network (IVN) where the nodes correspond to large-cap international stock markets while the links account for significant volatility lead-lags. Also included in the analysis is the International Trade Network (ITN), whose links measure bilateral export-import flows thus, capturing fundamental interconnections between countries. I find that the IVN and the ITN resemble each other closely pointing out that volatility does not disseminate randomly but tends to spread across fundamentally related economies. I also note that the amount of lagged volatility reactions in the IVN is consistent with the notion of gradual diffusion of information across investors who are subject to limited attention and home bias. I formally test this hypothesis by using as a direct proxy of investors’ attention the aggregate search frequency in Google. The empirical results support this intuition indicating that higher volatility surprises in key foreign markets predict higher domestic attention upon those markets in subsequent days. Once domestic attention is captured by such external shocks, it is contemporaneously transformed into higher domestic volatility.