2015年股市大跌前后的股指期现关系

发布时间：2018-01-15 22:11

本文关键词：2015年股市大跌前后的股指期现关系　出处：《上海社会科学院》2017年硕士论文　论文类型：学位论文

【摘要】：2015年6月至8月中国股市大跌期间,股指期货成交异常活跃,成交持仓比奇高,多个交易日出现相对现货指数的大绝对值负基差。很多人将股指期货视为股市大跌的重要推手,当年8月底和9月初监管层也出台了对股指期货交易的严厉限制措施。此后,关于股指期货在股市大跌中扮演的角色,业界和学术界有不少研究和争论。本文从股市大跌前后的股指期现关系——主要是期现日内收益率的领先滞后关系这一方面——入手切入这一问题。主要工作是使用1分钟高频数据,通过向量自回归和多元线性回归这两种简洁有效的方法,分析大跌前后股指期现收益率领先滞后关系的基本特征、前后变化以及在不同市场行情下的差异。结果表明,相比大跌之前和之后,股市大跌期间股指期货对现货的领先程度或者现货对期货的滞后程度保持不变或者有所减弱,而非显著增强;如果只考虑股市大跌期间,那么相比指数涨跌温和(用1分钟内指数涨跌幅度量)的时段,期货对现货的领先强度在指数大涨大跌时段明显增强,但这种增强在指数大涨和大跌两种情形下具有对称性,而不是在大跌时更强,即股指期货并没有单向的助跌作用。这些结果说明,从期现日内收益率动态关系的角度来看,股指期货导致股市大跌的说法缺乏依据。本文在两种方法原型的基础上进行了一些创新和改进:在使用向量自回归模型估计期现日内收益率在每个交易日的相互领先关系时,不仅对两个方向上是否有领先关系进行推断,还通过一系列系数联合显著性检验计算出每个交易日领先时间长度的上下界;进行多元线性回归时,不仅将现货收益率作为被解释变量,将期货收益率的领先、同期、滞后项作为解释变量以估计某个时间区间现货对期货的领先滞后关系并对误差项异方差和自相关予以修正,还通过在回归方程中添加虚拟变量的方法,评估同一时间区间内市场短期行情不同时这种关系是否有显著变化。本文的另一个创新之处是分析角度的多样化:4.3.2和4.3.3以交易日为单位分析期现关系的动态变化;4.4.2分三个时间段——股市大跌之前、股市大跌期间、股市大跌结束且股指期货限制措施实施后——分析股市大跌期间的期现关系与之前和之后的差别;4.4.3和4.4.4聚焦股市大跌期间,对比期现关系在指数涨跌幅不同或期现相对交易活跃度不同的时段的差异。不同分析角度得到的结果具有较好的一致性,表明本文结论是稳健的。本文基本结构是:前两章回顾国际和国内股指期货市场的起源和发展现状,并对相关研究进行综述;第三章概括了 2015年股市大跌期间现货和期货市场的异常表现,还对股指期货限制措施进行了归纳梳理;第四章先是简要论述选择1分钟抽样数据的理由,总结了数据基本特征,后面分别用两种方法进行统计推断,并对两种方法估计结果的含义差别进行了说明,最后列出后续可以进行的工作。
[Abstract]:From June 2015 to August Chinese during the stock market crash, the stock index futures traded positions were very active, beach high, a number of trading days appear relatively large stock index absolute value of negative basis. Many people will be an important promoter of stock index futures as the stock market crash, the same year the end of August and early September regulators have introduced strict restrictions on stock index futures the transaction of stock index futures. Then, play the role in the stock market crash, there are a lot of research and debate the industry and academia. The relationship between the stock of the current stock market before and after the main stage yields are now days the lead lag relationship in this aspect: to start cutting the problem. The main work is to use high frequency the data for 1 minutes, through vector auto regression and multivariate linear regression of these two kinds of simple and effective methods, analysis of the basic characteristics of the stock of the current rate of return fell after the lead lag relationship between before and after. The differences and changes in different market conditions. The results show that, compared to before and after the crash, the stock market during the period of stock index futures on the spot of the leading level or spot on the lag of futures remained unchanged or decreased, rather than increased; if only consider the stock market during the crash, then compared with moderate index change (index change amount within 1 minutes) time, leading strength on the spot futures index rose fell markedly in time, but the increase has symmetry in the index rose and fell under two cases, rather than more in the crash, the stock index futures is not a one-way or help role. These results suggest that the benefits from now the daily rate dynamic relations perspective, the stock index futures and stock market crash is the lack of basis. This paper makes some innovation and improvement in the two methods based on the prototype: Using the vector auto regression model to estimate the interaction relationship between the days leading yields on each trading day, not only to infer whether there is relationship between the two leading direction, is also calculated in each trading day ahead of time length on lower bound by a series of coefficients significant test; multiple linear regression, not only the spot rate of return as explanatory variables, the futures rate of return over the same period, lead lag as explanatory variables to estimate a time interval on the spot futures lead lag relationship and correct errors of heteroscedasticity and autocorrelation, by adding dummy variables in the regression equation, with short-term market assessment a time interval in the market and whether this relationship has changed significantly. Another innovation of this paper is to analyze the perspective of diversification: 4.3.2 and 4.3.3 to the trading day as a unit analysis The relationship between the dynamic change of 4.4.2; before three time periods: the stock market, the stock market crash, stock market and stock index futures ended after the implementation of restrictive measures -- Analysis of the stock market fell during the present relationship and difference before and after the period of 4.4.3 and 4.4.4; on the stock market fell, compared the present relationship of differences or different rose or the relative activity of different period in different angle analysis index. The results obtained are in good agreement, show that this conclusion is robust. The basic structure of this paper is: the origin and development of the review of the international and domestic stock index futures market in the first two chapters, and reviews related research; the third chapter the abnormal performance of spot and futures markets during the stock market crash of 2015, the stock index futures restrictions combed; the fourth chapter briefly discusses the choice of 1 minutes sampling The reason of data is summarized, the basic characteristics of data are summarized, and two methods are used for statistical inference. The meaning difference between the two methods is explained. Finally, the following works can be listed.

【学位授予单位】：上海社会科学院
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F724.5;F224

【相似文献】