基于不同风险度量的投资组合模型实证分析

The Empirically Analysis of Portfolio Optimization Models Based on Different Measure of Risk

利用上海证券市场的实际交易数据对半绝对离差模型,基于分位数的绝对离差模型(Ruszczynski&Vanderbei,2003)以及MV模型进行了实证比较.通过分析各模型的全局最小风险组合,计算发现,半绝对离差模型比其他模型具有更好的样本外业绩.

This paper compare portfolio optimization models based on different risk measurement using data from Shanghai securities market. We compares the global minimum risk portfolios of the different models：semi-absolute deviation model,weighted absolute deviation from the quantile model（ Ruszczynski ＆ Vanderbei,2003）and mean-variance model. The empirical study shows that semi-absolute deviation model is superior to the other two models and weighted absolute deviation from the quantile model is better than the mean-variance model.