成份股票价格指数编制模型的实证分析

【作者】 姜铁军；

【导师】 李从珠；

【作者基本信息】 北方工业大学 ， 数量经济学， 2001， 硕士

【摘要】 股票价格指数是反映股票市场价格波动的重要指标，不同的投资主体会从不同角度 看待市场指数，从自身需要出发使用不同类型的指数，因此在发达国家股票市场中存在 着各种用途的股价指数。建立一个完善的市场指数体系可以为指数基金提供投资评价的 工具，为金融投资者提供品种齐全的现代金融衍生产品，为金融避险者提供头寸对冲工 具，为股市分析家提供研究判断股市的动态信息。 由于目前沪、深两市己经发布的成份指数均属于窄基指数，样本容量和市值比重较 小，个别股票在样本股票中权重过大，难以涵盖更多的代表性股票。国外的研究表明宽 基指数比窄基指数在预测经济周期转折点方面更为有效，同时也更受指数基金和股指期 货投资者的青睐。 针对现有股票指数种类单一的缺陷，本文通过因子分析、聚类分析、相关分析等数 理统计方法，在假定沪、深两个股票市场融合为一个市场的前提下，实证研究了编制成 份股票价格指数的选股模型，使得该模型在最大程度上利用了上市公司的财务报告以及 股票价格反映出的股票波动特性，保证样本股票覆盖到现有的各个行业，同时使该选股 方法能够适应未来股市扩容发展的需要。 本文的目的就是提出一个能够相对客观、全面地反映市场的宽基成份指数的编制模 型。使通过该模型选出的成份股票指数对于指数基金而言，具有更高的可操作性和更低 的跟踪成本；而对于市场中的恶意炒做者而言，具有更高的炒做成本和难度。 论文主体共分为四个部分。在第一章绪论中主要介绍了世界各国编制的主要股票价格指数的一般原则，并讨论 了当前国内在股市指数编制方面做的较好的几种成份股票价格指数的编制方法，它们成 为后续部分成份指数（S362）编制过程的参照。其余三章内容均围绕绪论中给出的研究 思路和方法进行实证分析。 在第二章中首先分析了成份股票股本结构变动对股指计算的影响，以及进行基期调 整的各种可能情况。同时还进行了以总股本和以流通股本加权的优劣分析，两种不同情况下指数计算机期修正方法的差异。 解决了指数计算、修正方法问题之后，就面临成份指数样本股的选取模型设计问题了，这是第三章主要讨论的内容，也是本文的研究重点。我们把样本股票的选择定位在反映市场各类特色股票变化的基础上，并提出了六条选股原则。基于上述原则建立了三重划分的选股模型，即：行业划分、波动特性划分、经营业绩划分。结合流通市值的大小最终选出了362只成份股票样本。最后给出了股票样本的调整原则和周期。 第四章单独用来分析所算成份指数（5362）半年来的运行特征。并从市场代表性、相关性、投资收益和风险特性以及BETA值特征四个方面对该指数进行检验，论述了通过该模型所选股票具有的优良特性。更多还原

【Abstract】 Stock price index is of great important indicator to reflect the volatility of the stock price in the security market. Investors?attitude towards the stock indices varies widely and they often use different indices to meet their investment demands. Therefore, diversified series of indices with different functions coexist in the security market of developed country. A perfect stock index system can provide evaluation tools of investment for index funds, various derivatives for financial investors, measures for risk aversions to reduce their position risk as well as dynamic information for security market analyst.SSSPresently, most published indices belong to small size index with the characters of low liquidity, inappropriate industry coverage and weight, high cross-ownership. Some research papers show that indices with large size are more sensitive to the turning points of economic then that with small size. Index funds and investors of stock index future prefer using the former index as trade object.On the assumption of the unification of two stock markets, we apply several commonly used statistical methods, such as factor analysis, clustering analysis, correlation analysis, to build an effective model by which an objective stock sample can be obtained for computing index. Meanwhile, the information derived from the accounting annals of public owned company and the idiosyncrasy revealed by fluctuant pattern of stock price have been fully exploited. The index computed by this model covers all kinds of industries and regions and it can also be used in the further developed security market.The main purpose of this dissertation is to bring forward an objective and comprehensive model by which we can calculate a large size index that can offer more convenience and low tracing cost for risk management.The dissertation is divided into four chapters.Chapter 1 provides a general introduction to the methodology and subject of indices and a brief treatment of basic concepts of index. It also contains the research scheme and empirical analysis plan.Chapter 2 presents general views on the difference between using total number of shares and liquid number of shares as the weight of index. An empirical study is also conducted to show the real difference. Finally, the algorithm of index calculation and base date adjustment is given.Chapter 3 deals with the illustration of stock selection model and the principles of adjustment, specifically the 5362 stock index selection model.Chapter 4 demonstrates the outstanding characteristic of S362 index from four aspects: a representative aggregation of stocks in security market, high correlation with main stock indices, high investment return and low systematic risk, same BETA value distribution as stock market.更多还原