基于非参数估计框架的期望效用最大化最优投资组合

发布时间：2018-08-20 19:48

【摘要】：本文基于期望效用最大化和非参数估计框架研究了最优投资组合选择问题。和以往大多文献假定资产收益率服从某些特定分布不同资产收益率的分布类型无需作任何假设。首先在一般效用函数下,利用组合收益率密度函数的非参数核估计给出了期望效用的基本非参数估计公式,并建立了期望效用最大化投资组合选择问题的基本框架。然后,在投资者具有幂效用函数的假定下,给出了期望效用具体的非参数计算公式,并给出了求解最大期望效用的数值算法。最后,利用中国证券交易所11支股票日收益率的真实数据给出了一个数值算例。本文提出的非参数估计框架具有一般性,还可以进一步用来研究各种现实条件下(如各种现实不等式约束和具有交易成本)的投资组合管理问题。
[Abstract]:In this paper, the optimal portfolio selection problem is studied based on the framework of expected utility maximization and nonparametric estimation. And most of the previous literatures assume that there is no need to make any assumptions about the distribution types of asset returns from certain specific distributions. Under the general utility function, the basic nonparametric estimation formula of expected utility is given by using the nonparametric kernel estimation of the portfolio return density function, and the basic framework of the portfolio selection problem of maximizing expected utility is established. Then, under the assumption that the investor has a power utility function, the non-parametric formula for calculating the expected utility is given, and the numerical algorithm for solving the maximum expected utility is given. Finally, a numerical example is given using the real data of the daily return rate of 11 stocks on the China Stock Exchange. The nonparametric estimation framework proposed in this paper is general and can be further used to study portfolio management problems under various real conditions such as real inequality constraints and transaction costs.
【作者单位】：广东外语外贸大学信息学院;中山大学管理学院;
【基金】：国家自然科学基金重点项目(71231008) 广东省高等学校高层次人才项目广东省自然科学基金项目(S2011010005503) 广东省高等院校科技创新项目(2012KJCX0050) 广东省科技计划项目(2012B040305009) “全国统计科学”研究计划一般项目(2013LY101)
【分类号】：F830.59;F224

【共引文献】