基于时间加权历史模拟法的VaR来构建最优投资组合

Constructing the Optimal Portfolio Based on VaR of Time-Weighted History Simulation Method

摘要: 假设收益率服从正态分布时，均值–方差模型常被用于构建最优投资组合。但很多情况下，收益率并不服从正态分布。本文首先构造股票投资价值的衡量指标，根据指标对股票的优劣进行排序；然后利用时间加权历史模拟法来计算投资组合的VaR，建立相应的均值-VaR模型；最后利用均值-VaR模型构建中国股票市场的最优投资组合，预测最优投资组合的风险。此方法可有效避免收益率服从正态分布的假定。

Abstract: On the assumption that yields obey the normal distribution, mean-variance model is frequently used in the optimal portfolio; but in many cases, yields don’t obey the normal distribution. Firstly, we construct a measure index of stock investment value and sort the merits of stock by the index. Then the VaR of portfolio is calculated by using the time-weighted history simulation method and Mean-VaR model is built accordingly. Finally, the optimal portfolio of Chinese stock market is con-structed by using the Mean-VaR model, and the risk of optimal portfolio is predicted. The assump-tion of normal distribution can be avoided effectively by using this method.