Does idiosyncratic volatility matter at the global level?

Abstract

I test the existence of a time-series relationship between the aggregate idiosyncratic volatility and the market index return at the global level by introducing various global measures of aggregate idiosyncratic volatility. I offer four definitions of aggregate global idiosyncratic volatility (GIVOL) based on factor models and two other definitions which are free from factor models. Regardless of whether I use model-dependent or model-independent measures I find no evidence of a robust and significant relation between the aggregate GIVOL and the global market return. This result is valid for four different sub-periods and four different subsamples reflecting the different states of the economy and the stock market. It is also robust to the inclusion of several control variables. As global idiosyncratic volatility is not a priced factor in the intertemporal asset pricing framework the results indicate that international diversification is still effective in eliminating idiosyncratic volatility despite the globalization process.