指数投资与跟踪复制的实证研究

发布时间：2018-01-01 00:38

本文关键词：指数投资与跟踪复制的实证研究　出处：《南京大学》2017年硕士论文　论文类型：学位论文

【摘要】：指数投资是被动投资的主要形式之一,是指以某一市场指数为标的进行复制跟踪的投资行为,以期获得与标的指数相近的投资收益,具有风险小、费用低、流动性高的优点。随着2015年A股市场股灾的爆发,大量主动管理型产品产生巨幅回撤,使投资者们愈发意识到以指数投资为首的被动投资方式的重要性。越来越多的私募基金开始涉足指数跟踪、指数增强与绝对收益类产品,但受限于信息、规模与操作成本约束,需要面对指数精确权重数据费用偏高、成分股中零散股比例偏高、实盘交易操作成本偏高等现实约束,很难以较低的操作复杂度与交易成本实现精确的全样本指数复制。本文应此背景产生,目的是找到一种以公开数据为基础、以较低的交易频率与较少的持仓股票数量实现较低跟踪偏离度的指数抽样复制方法,为中小型指数化产品提供具有实盘指导价值的指数投资策略。研究过程中,综合运用了理论分析法和实证分析法。首先,通过文献综述法归纳分析现有研究成果,确定指数投资的理论基础和实证方法。随后,选择与市场中实际投资行为一致的指数作为标的,通过蒙特卡洛模拟法指出经典跟踪偏离度算法的问题并得出改进方案。最后,基于实践中具备可行性的投资行为假设,对指数跟踪复制组合的跟踪偏离度进行实证测试。使用控制变量法建立实验对照组,对不同指数抽样复制方法的实证效果进行排序,以便分析不同抽样复制方法的优劣。在实证方面,选择国内股票市场中最具影响力的沪深300指数与中证500指数作为研究对象,基于2006年至2016年共计11年的数据进行实证检验。根据逻辑顺序,首先对基于指数月末公开权重数据进行调仓的全样本复制法进行实证检查,证明仅以公开数据与较低的交易频率进行指数复制是可行的。随后,引入持仓股票数量约束,通过对分别简单随机抽样复制法、最大权重抽样复制法、最小权重抽样复制法、行业分层抽样复制法及其组合的实证对比分析,并针对每种抽样方法计算在各种抽样比例下的跟踪偏离度。最后,汇总实证数据,绘制以抽样复制比例为横坐标、以跟踪偏离度为纵坐标并将多种抽样复制方法结果同列的图表,直观地得到了不同抽样复制方法之间的优劣关系,以及抽样复制比例与跟踪偏离度之间的非线性关系。最终,本文得出使用最大权重复制法叠加行业分层抽样的指数跟踪复制效果最好的结论,并进一步发现了沪深300跟踪复制组合中更依赖大权重成份股的特征。在成果转化方面,本文的实证结果可为指数投资提供具体建议。首先,本文通过对月末全样本复制的实证检验,证明以公开数据与低频调仓,通过全样本复制即可实现较好的跟踪复制效果,为受信息获取与交易操作约束但不受规模约束的指数化产品提供了简单可行的指数复制方案。其次,针对更一般的情况,将规模约束纳入考虑后,当必须对指数进行抽样复制之时,则最大权重复制法叠加行业分层抽样的跟踪复制效果最好,并且可以基于抽样复制比例与跟踪偏离度之间的增量关系,确定最经济的抽样复制比例。
[Abstract]:Index investment is one of the main forms of passive investment, refers to a market index as the underlying replicated investment behavior tracking, in order to obtain the similar index of investment income, has low risk, low cost, the advantages of the high mobility of A stock market crash of 2015. With the outbreak of a large number of active management type of products produced huge retracement, so that investors are increasingly aware of the importance of passive index investment investment led. More and more private equity funds began to get involved in index tracking, enhanced index products and absolute income, but limited information, scale and operation cost constraints, need to face the index weight data accurate cost is high, stocks in the high proportion of scattered shares, trading operation of high cost and other practical constraints, it is difficult to lower operation complexity and transaction costs to achieve full sample index accurately reproduced. This should be the The background, the purpose is to find an open data base, to the number of trading frequency and lower stock positions to achieve low tracking deviation index method of sampling replication, provide the index investment strategy has a guiding value for small and medium sized index product. In the course of the study, the integrated use of the theoretical analysis and empirical analysis. Firstly, through the literature review summarized the existing research results, the theoretical basis and empirical method of index investment. Then, select the market and the actual investment behavior consistent index as a subject, it is pointed out that the classical tracking deviation algorithm and obtains the improvement scheme by Monte Carlo simulation method. Finally, with the assumption that the feasibility of investment behavior in practice based on the index tracking portfolio replication tracking deviation of the empirical test. The establishment of the experimental group used as control variable method, To sort the positive effect of different index sampling replication method, in order to analyze the different sampling replication methods. In the empirical analysis, select the Shanghai and Shenzhen's most influential domestic stock market in 300 index and the CSI 500 index as the object of study, an empirical test was conducted from 2006 to 2016 a total of 11 years. Based on the data according to the logical order, first of all at the end of the index weight based on the full sample data open copy method to adjust positions of empirical examination, that is only based on the public data and transaction frequency is low the index replication is feasible. Then, the number of constraints on stock positions, respectively by simple random sampling replication method, maximum weighted sampling replication method, the minimum weight of sample copy the empirical analysis method, comparative method and stratified sampling replication industry portfolio, and for each sampling method in the calculation of various sampling rate tracking deviation . at last, summarize the empirical data, to draw a sample copy of the proportion as the abscissa and ordinate to track deviation degree and a variety of sampling replication method results with the column chart, obtained the advantages and disadvantages between different sampling methods, sampling and replication ratio and the relationship between the degree of deviation from the nonlinear tracking. Finally, we use maximum weight stratified sampling replication method of superposition industry index tracking copy the best effect, and further found that the Shanghai and Shenzhen 300 tracking copy feature more dependent on weight stocks portfolio. In the transformation, the empirical results of this paper can provide specific suggestions for the investment index. First of all, based on the empirical test at the end of the whole sample copy, proof by the public data and low frequency transfer positions, through the whole sample copy can achieve better tracking effect for replication, access to information and trading The operating constraints, but not by index product scale constraints provides a simple and feasible index replication scheme. Secondly, according to the more general case, the scale constraints into consideration, when the copy must be sampled on the index, the maximum weight of copy copy effect tracking method overlay industry stratified sampling, and sampling the proportion of copying and tracking based on the relationship between the degree of deviation increment, determine the proportion sample copy of the most economic.

【学位授予单位】：南京大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.51

【参考文献】