Optimal and naive diversification in an emerging market: Evidence from China's A-shares market

This paper empirically investigates the out-of-sample performance of the 1/N naive rule and the Markowitz mean–variance strategies in the largest emerging market (i.e., China's A-shares market) and provides three new findings. First, we show that some mean–variance optimization strategies can outperform the 1/N rule in China's A-shares market, while minimum-variance strategies cannot. Using certainty equivalent return (CER) instead of Sharpe ratios does not change our results qualitatively. Second, we find an obvious advantage of mean–variance optimization when N is large. Third, when transaction costs are taken into account, the profitability of the unconstrained mean–variance optimizations almost vanishes, while the profitability of the mean–variance optimizations with the short-sale constraint remains. Our results are robust to using a shorter estimation window of about 60 months. These results provide support for the use of optimal diversification strategies in emerging markets.