Portfolios resampling and international diversification : a non-parametric stochastic dominance approach

This paper examines the impact of estimation errors on the financial portfolios optimization processes and investigates the controversy problem of the international and domestic optimal diversification strategies choice from an American investor’s point of view. We introduce the concept of portfolio resampling method and we use the nonparametric stochastic dominance approach based simulated p-values to define an optimal diversification choice. Estimation errors visualization shows that changes in input parameters imply large changes in portfolio composition and reveals considerably modification of MV efficient frontiers shape. The findings show that there exists substantial evidence of the international global diversification benefits. Risk-adverse American investor with an increasing utility function prefers the global international resampled diversification strategy. We find that domestic diversification beats only international major and emerging markets diversification. Dominance relationships between the entirely diversification strategies change according to the risk-aversion coefficient.