Edinburgh Napier University

  The author investigated the existence and significance of a global cross-sectional relation between idiosyncratic volatility and expected returns by introducing a global idiosyncratic volatility measure and globally diversified test assets. He found that portfolios with the highest and lowest global idiosyncratic volatility do not earn significantly different average returns, indicating no link between global idiosyncratic volatility and expected returns. His results show that global diversification is effective in stabilizing the returns of global test assets and that benefits from global diversification can be gained by diversifying across either countries or industries.

In this study, I investigated the existence and significance of a cross-sectional relationship between global idiosyncratic volatility and expected return. This study extends the debate on whether idiosyncratic volatility matters for expected returns at the global level by taking the perspective of a global investor. I introduced a global idiosyncratic volatility measure that is defined as the residual volatility in which the residuals are obtained by regressing the returns of globally diversified test assets on the global systematic risk factors in an international asset pricing framework. I formed global test assets from characteristic-sorted local supersector, local stock market, and global sector indexes. The use of characteristic-sorted indexes facilitates diversification at the domestic, international, and industrial levels and thus leads to the formation of globally well-diversified test assets. I sorted the test assets on the basis of global idiosyncratic volatility and formed three portfolios. Portfolio 1 consisted of the test assets with the lowest global idiosyncratic volatility; Portfolio 3 contained the test assets with the highest global idiosyncratic volatility. I performed average return difference tests for Portfolios 3 and 1 to see whether a relationship exists between global idiosyncratic volatility and expected return. I consistently found no statistically significant return difference between the highest- and lowest-GIVOL portfolios and thus found no evidence of a relationship between global idiosyncratic volatility and expected return.