The myth of international diversification

We test the proposition that international diversification is effective in reducing risk. The traditional underlying argument is that low correlations of international stock returns make the variance of an international portfolio lower than the variance of a purely domestic portfolio when long positions are taken on the domestic and foreign markets. Our analysis of more than 100 portfolios involving developed and emerging markets shows that correlations are not adequately low to produce effective diversification when long positions are taken. In a few cases involving developed markets only, correlations are high to the extent that taking opposite positions (long and short) produces effective diversification. The results cast serious doubt on the effectiveness of international diversification in reducing risk.