基于分层PCA的指数跟踪及实证

发布时间：2017-12-27 13:36

本文关键词：基于分层PCA的指数跟踪及实证　出处：《中国管理科学》2013年S1期 　论文类型：期刊论文

【摘要】：在当前金融市场中,被动投资策略具有操作简便、成本低、风险低等一系列的优点,因此被越来越多的投资者所青睐。然而,被动投资策略的好坏主要通过跟踪误差来评判,如果跟踪误差较大,就不能够为投资者提供接近标的指数的收益。基于此本文提出了分层主成分分析方法,采用完全复制标的指数的投资策略来构建因素投资组合选择模型最小化跟踪误差,在此基础上通过改变分层的数量控制投资组合的跟踪误差。实证结果表明,基于本文提出的分层主成分分析方法,样本内与样本外的跟踪误差都很小,而且随着样本内层数的增加,样本外的跟踪误差有很大改善。
[Abstract]:In the current financial market, passive investment strategy has a series of advantages such as simple operation, low cost and low risk. Therefore, it is favored by more and more investors. However, the quality of passive investment strategy is mainly judged by tracking error. If the tracking error is large, it can not provide investors close to the index's income. Based on this, a hierarchical principal component analysis method is proposed in this paper, which uses the investment strategy of completely replicating the index index to build a factor portfolio selection model to minimize tracking error. On this basis, we can control the tracking error of portfolio by changing the number of layering. The empirical results show that, based on the hierarchical principal component analysis proposed in this paper, the tracking errors between samples inside and outside samples are very small, and with the increase of the number of samples, the tracking error outside the sample has been greatly improved.
【作者单位】：中国科学院数学与系统科学研究院;
【分类号】：F224;F830
【正文快照】： ^ S 指数跟踪技术是被基金公司广泛应用在资产管理上的一种被动投资策略,相对于主动投资策略而言,具有投资风险低、管理成本低和长期有较好收益等优点,被越来越多的基金公司所采用。所谓的指数跟踪技术是指通过购买全部或者部分标的指数成份股,来完全复制或者部分复制跟踪标的

【参考文献】