沪深300期指和指数的相关性及交易策略探究

发布时间：2018-01-23 06:52

本文关键词： 沪深300期指长期关系因果关系交易策略　出处：《哈尔滨工业大学》2014年硕士论文　论文类型：学位论文

【摘要】：金融业快速发展,风险的不确定性增加,各种金融衍生工具创新层出不穷。期指被认为是最为成功的金融创新,是世界各国期货市场交易最活跃的金融衍生工具之一。期指具有价格发现、套利、套期保值和稳定证券市场等作用,深受投资者和决策者的青睐,而期指的高杠杆性又使市场存在潜在的风险。为了完善中国资本市场体制,中国金融期货交易所在2010年4月16日正式发布沪深300期指并上市交易,但是当日沪深300股票指数大幅度下跌,这让投资者怀疑期指是否具有稳定市场和价格发现的功能,因此有必要探究沪深300期指和指数之间的相关性,并且在此基础上提出合理的交易策略。本文选取从2010年4月16日到2011年4月15日的一分钟沪深300期指和指数高频数据作为样本数据,研究有以下主要结论:(1)为了方便计算和减少误差,对沪深300期指和指数取自然对数log,并且通过单位根检验,发现沪深300期指和指数是一阶单整的,说明log差分一阶意味着收益率是平稳的,并且可以采用时间序列方法对其进行预测。(2)利用门限自回归模型分析得出沪深300期指和指数具有长期关系,并通过引入有效能参数改进门限自回归模型,有效能参数在一定程度上显示了平稳趋势,更能够反映出两者的长期关系,分析结果发现沪深300期指和指数同样具有长期关系。在两者具有长期关系的条件下,采用门限误差修正模型对两者的因果关系进行了探究,其中指示函数采用有效能参数,发现沪深300期指对股票指数有较强的引导能力,领先股票指数15~26分钟之间,而股票指数对期指却有较弱的引导能力,领先期指8~47分钟。(3)基于上述分析结论,本文对沪深300期指进行了预测,选取样本为2010年4月16日到2011年4月15日作为样本内数据,比较了自回归模型、移动平均模型和差分自回归移动平均模型的预测精度,对比发现移动平均模型要比自回归模型和差分自回归移动平均模型精确度要高,但是差别不是很大。文章还比较了样本时间段为2011年4月16日到2013年4月9日的1分钟沪深300期指的预测数据和滞后阶数18期,20期,22期和24期后的股票指数的走势,结果发现滞后22期股票指数和期指的预测数据具有相似的走势,滞后24期的股票指数和期指的真实数据具有相似的走势,因此投资者可以通过沪深300期指的走势来判断预测股票指数的走势,提前做好交易决策,从而获取收益。
[Abstract]:With the rapid development of the financial industry and the increase of the uncertainty of the risk, various kinds of financial derivatives innovation emerge in endlessly. The index index is considered to be the most successful financial innovation. Futures index is one of the most active financial derivatives in futures markets in the world. The futures index has the functions of price discovery arbitrage hedging and stabilizing the securities market and is favored by investors and policy makers. In order to perfect China's capital market system, China Financial Futures Exchange officially released the Shanghai and Shenzhen 300 futures index and listed on April 16th 2010. However, the Shanghai and Shenzhen 300 stock index fell sharply on the same day, which made investors doubt whether the futures index has the function of stabilizing the market and price discovery, so it is necessary to explore the correlation between the Shanghai and Shenzhen 300 index and the index. And on this basis put forward a reasonable trading strategy. This paper selects one minute Shanghai and Shenzhen 300 index and index high frequency data from April 16th 2010 to April 15th 2011 as sample data. The main conclusions of the study are as follows: (1) in order to facilitate the calculation and reduce the error, the natural logarithm is taken for the index and index of the Shanghai and Shenzhen 300 futures index, and the unit root test is adopted. It is found that the index and index of Shanghai and Shenzhen 300 are one-order and single-integral, indicating that the first order of log difference means that the yield is stable. And we can use time series method to predict it. 2) using threshold autoregressive model analysis, we can conclude that the index of Shanghai and Shenzhen 300 has a long-term relationship. And by introducing effective parameters to improve threshold autoregressive model, the effective parameters to a certain extent show a stable trend, and can reflect the long-term relationship between the two. The results show that the CSI 300 index and the index also have a long-term relationship. Under the condition of long-term relationship between the two indexes, the causal relationship between them is explored by using the threshold error correction model. The indicator function adopts effective parameters, and it is found that the CSI 300 index has a strong ability to guide the stock index, leading the stock index between 15 minutes and 26 minutes. On the other hand, the stock index has a weak ability to guide the futures index, leading the futures index 847 minutes. 3) based on the above conclusions, this paper forecasts the Shanghai and Shenzhen 300 futures index. From April 16th 2010 to April 15th 2011, the prediction accuracy of autoregressive model, moving average model and differential autoregressive moving average model were compared. It is found that the moving average model is more accurate than the autoregressive model and the differential autoregressive moving average model. But the difference is not very big. The paper also compares the prediction data of 1 minute CSI 300 index from April 16th 2011 to April 9th 2013 and the lag order of 18 periods and 20 periods. The trend of the stock index after 22 period and 24 period. The result shows that the forecast data of the stock index and futures index that lag behind 22 periods have similar trend. The real data of stock index and futures index with 24 periods lag have similar trend, so investors can judge and forecast the trend of stock index through the trend of Shanghai and Shenzhen 300 futures index, and make a good trading decision in advance. In order to gain income.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F724.5;O212.1

【参考文献】