论文部分内容阅读
本文以从沪深300指数中选取的四只行业指数为样本,采用样本协方差矩阵、单指数模型矩阵、常量相关矩阵、数量矩阵和两参数模型矩阵预测方法,结合多周期M-V投资策略,探讨了不同的协方差矩阵预测方法对多周期M-V最优投资组合模型产生的影响。实证研究表明,在多周期M-V投资决策下,无论是基于有效边界、还是基于不同的风险厌恶系数、抑或是基于不同的多周期角度来看,如果投资组合中不存在无风险资产,采用常量相关矩阵可以得到最优的投资策略;如果投资组合中存在无风险资产,采用数量矩阵可以得到最优的投资策略。而无论投资组合中是否存在无风险资产,样本协方差矩阵得到的投资策略的效果基本上都是最差的。
In this paper, four industry indices selected from the CSI 300 Index are used as samples. The sample covariance matrix, single exponential model matrix, constant correlation matrix, quantity matrix and two-parameter matrix forecasting method are used in combination with the multi-cycle MV investment strategy The influence of different covariance matrix prediction methods on multi-period MV optimal portfolio model. Empirical studies show that under the multi-cycle MV investment decision-making, whether based on the effective border or based on different risk aversion coefficients, or based on different multi-cycle perspective, if there is no risk-free assets in the portfolio, The matrix can get the optimal investment strategy. If there is risk-free asset in the portfolio, the optimal investment strategy can be obtained by using the quantitative matrix. Regardless of whether there are risk-free assets in the portfolio, the effect of the investment strategy obtained by the sample covariance matrix is basically the worst.