The Research On Construction Of Index Proxy Portfolio

Markowitz's mean-variance model indicates that the optimum risk asset being the market portfolio. But there are so many stocks in the market but investors with only limited capital that it is not practical to invest in the market portfolio. Thereby, as an efficient way of dispersing risk and obtaining market-average income, investing according to the index is warmly welcomed by numerous investors as stock market efficiency improving everyday.This thesis focuses on index proxy portfolio, which can decrease portfolio volume and reduce the capital requirement of investor, in the mean time, closely track market index. On the basis of research on market efficiency, index funding, clustering algorithm and time factor, 295 stocks in Shenzhen stock market are selected as the research objects. Clustering algorithm with time factor is applied to choose portfolio population, and then Single Index Model is used to calculate the weight of every individual stock in order to construct an index proxy portfolio to track Shenzhen Composite Index. We assume that the investor adopts buy and hold policy, then a series of index are used to scale the trace process in the following three months. The demonstration shows that clustering algorithm is good for selecting handful stocks from the whole market and appropriate time factor can improve tracking result significantly.The most important provable innovation of this thesis lies in the introduction of time factor. Comparing the impact on index proxy portfolio by different time factor, the result comes that time factor in exponential form can describe the actual market best.