【摘要】 本文探讨沿如下新的思路构造投资组合来最有效地单向"跟踪"某个金融指数:为灵活刻画不同投资者对跟踪风险的感知,我们采用跟踪偏差相应阶数的下半矩作为风险的度量;为较好描述证券收益分布的厚尾现象,我们假设市场上风险资产的收益服从多元t分布.在分析基于上述框架所建立跟踪优化模型基本性质的同时,我们还设计出相应的求解算法.实证检验结果不仅表明新模型与算法的有效性和实用价值,而且较好地克服了基于MAD等准则的传统跟踪方法的不足.更多还原

【Abstract】 The problem of efficiently and one-sidedly"tracking"some financial index through constructing portfolios is investigated in this paper from the following new aspects:in order to flexibly describe different investors’ perception to the tracking risk,the lower partial mo- ments of the tracking deviation is adopted as the risk measure;to better reflect the fat-tail phenomenon of security return distributions,it is assumed that the joint return distribution of stocks follows the multivariable t-distribution.After analyzing fundamental properties of the established tracking optimization problem,we develop a numerical algorithm for solving the optimization problem.Empirical results demonstrate the efficiency and practical value of our new tracking model and the corresponding algorithm,which also overcome shortcomings of traditional tracking methods such as the MAD model.更多还原