基于相关性分析的中国金融市场演化特性研究

发布时间：2018-07-03 05:23

本文选题：偏相关性分析 + 秩相关系数　；参考：《南昌航空大学》2017年硕士论文

【摘要】：金融市场作为当今经济领域的一个组成部分,在人们的生活中扮演着重要的角色,金融理论和实证研究引起了来自不同领域科学家的兴趣。本文基于中国金融市场的实证数据,运用了不同的相关性分析方法,研究了中国金融市场的演化特性,主要研究工作和成果如下:(1)基于偏相关性分析,研究了多种因素对中国金融市场金融资产间相关性的影响。首先,通过对比股票之间的皮尔逊相关性与剔除国内市场指数影响的股票之间的偏相关性,研究国内市场指数对中国金融市场的影响;其次,通过对比剔除国内市场指数影响的股票之间的偏相关性和剔除国内市场指数、个股影响的股票之间扩展的偏相关性,研究了个股对中国金融市场的影响;然后,基于证监会行业(CSRC)分类对股票的划分,研究行业对个股的影响及各行业间的关联性;最后,通过对比剔除国内市场指数影响的股票之间的偏相关性和剔除国内市场指数、国外市场指数影响的股票之间扩展的偏相关性,研究国外市场指数对中国金融市场的影响。研究结果表明,国内市场指数对国内市场股票的相关性影响很大,而美国、欧洲和亚洲等外部市场指数对国内市场股票的影响则很微弱。股票会受到不同行业的影响,但主要影响来自于本身所属行业或密切相关行业。金融、保险行业与其他行业的相关性较弱。(2)基于秩相关性分析,研究了中国股票市场的结构演化特性。首先,将股票的时间分别以月、季度为单位划分;其次,根据时间的划分求出每个时间段股票市场之间的Kendallτ秩相关系数,其中τ系数用来描述股票市场的结构相似性;然后,通过计算系数τ与价格收益率的皮尔逊关联系数,研究市场结构相似性和价格收益率的关系;最后,研究系数τ随时间的变化趋势,研究股票市场结构的持续性。研究结果表明,市场结构相似性与大多数时间的价格收益率呈负相关,并且市场结构相似性随时间间隔增长而迅速衰减。(3)研究了中国股票市场的地理特性。首先,通过对股票每日的数据进行主成分分析,研究各省级行政区在中国股票市场的重要性;其次,研究基于位置参数的相关性;最后,研究距离分布特性和基于距离参数的相关性。研究结果表明,除了金融大波动时的深圳市场,股票位置对金融市场演化均有影响;股市中的距离呈两段式分布,短距离的概率远大于长距离的概率。上海证券交易所股票的相关性随着股票距离增大呈现弱衰减趋势,而深圳证券交易所股票的相关性随着股票距离增大基本保持稳定。综上所述,本文通过对真实金融市场的演化特征进行研究,能够为金融系统的风险控制提供一定的理论依据。
[Abstract]:As an integral part of today's economic field, financial market plays an important role in people's life. Financial theory and empirical research have aroused the interest of scientists from different fields. Based on the empirical data of Chinese financial market, this paper uses different correlation analysis methods to study the evolution characteristics of Chinese financial market. The main research work and results are as follows: (1) based on partial correlation analysis, This paper studies the influence of various factors on the correlation of financial assets in Chinese financial market. First of all, by comparing the Pearson correlation between stocks and the partial correlation between stocks excluding the influence of domestic market index, we study the impact of domestic market index on Chinese financial market. By comparing the partial correlation between stocks that exclude the influence of domestic market index and the extended partial correlation between stocks affected by individual stock, this paper studies the influence of individual stock on Chinese financial market. Based on the classification of CSRC to stocks, this paper studies the influence of industry on individual stock and the relevance of each industry. Finally, by comparing the partial correlation between stocks with excluding the influence of domestic market index and eliminating the domestic market index, This paper studies the influence of foreign market index on Chinese financial market. The results show that the domestic market index has a great impact on the domestic stock market, while the external market index, such as the United States, Europe and Asia, has a weak impact on the domestic stock market. Stocks will be affected by different industries, but mainly from their own industries or closely related industries. The correlation between finance and insurance industry and other industries is weak. (2) based on rank correlation analysis, the structural evolution characteristics of Chinese stock market are studied. Firstly, the time of stock is divided into month and quarter respectively. Secondly, the Kendall 蟿 rank correlation coefficient of stock market in each time period is obtained according to the division of time, where 蟿 coefficient is used to describe the structural similarity of stock market. By calculating the Pearson correlation coefficient between coefficient 蟿 and price return, the relationship between market structure similarity and price return is studied. Finally, the variation trend of coefficient 蟿 with time is studied to study the persistence of stock market structure. The results show that the similarity of market structure is negatively correlated with the price return in most time, and the similarity of market structure decreases rapidly with the increase of time interval. (3) the geographical characteristics of Chinese stock market are studied. Firstly, through the principal component analysis of the daily stock data, the importance of the provincial administrative regions in the Chinese stock market is studied. Secondly, the correlation based on the location parameters is studied. The distance distribution characteristics and the correlation based on distance parameters are studied. The results show that, except in Shenzhen market, stock position has influence on the evolution of financial market, and the distance in stock market is distributed in two sections, and the probability of short distance is much larger than that of long distance. The correlation of Shanghai Stock Exchange stock shows a weak attenuation trend with the increase of stock distance, while the correlation of Shenzhen Stock Exchange stock keeps stable with the increase of stock distance. To sum up, this paper studies the evolution characteristics of real financial market, which can provide some theoretical basis for the risk control of financial system.
【学位授予单位】：南昌航空大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.5

【参考文献】