中国股指期货收益率波动性与交易量、持仓量的关系探究

发布时间：2018-03-01 04:01

本文关键词： 股指期货持仓量交易量波动性　出处：《复旦大学》2013年硕士论文　论文类型：学位论文

【摘要】：为了研究股指期货市场收益率、持仓量和交易量之间的关系,本文通过研究沪深300期货合约,主力合约和月连续合约从2010年4月16日到2013年3月12日的样本数据,用前660个样本进行模型评价,用最后43个样本做样本外评价。首先本文使用GARCH模型,分别研究了持仓量和交易量对波动性的影响。研究结果发现,对于主力合约来说,交易量的滞后项对收益率波动性的影响很显著,并且交易量的滞后项与收益率波动性之间的关系是正向的。而持仓量的滞后项对于收益率波动性的影响也是非常显著的,并且持仓量的滞后项与收益率波动性之间的关系是负向的。这一结论与主流的研究类似。对于非主力合约,也就是我们所选取的月连续合约样本来说,GARCH模型回归的结果显示,无论是交易量,还是持仓量,与收益率波动性之间的关系都不显著。然后,本文使用格兰杰因果检验,检验三者之间的因果关系。格兰杰因果检验结果显示,交易量和波动性之间存在双向的格兰杰因果关系,持仓量是波动性之间的格兰杰原因,波动性并不是持仓量的格兰杰原因。交易量和持仓量之间具有双向的格兰杰因果关系。为了考虑这三个变量总体的相互影响,建立VAR模型,考量这三个变量整体的关系。主力合约VAR模型实证结果显示,从模型中,我们依然可以得到交易量的滞后项对波动性具有正影响,持仓量的滞后项对收益率波动性具有负的影响,持仓量和交易量之间存在很强的互相影响。而滞后期的波动性对持仓量没有影响。从月连续合约的VAR模型结果,本文发现,对波动率主要受到自身的影响,持仓量和交易量的影响比较弱。而持仓量和交易量之间的相互影响仍旧很强烈,波动性的滞后项对交易量也有显著的影响。脉冲响应分析发现主力合约和非主力合约的结论类似,波动率对自身有很强的影响,但这种影响的很短暂。交易量对收益率波动性的影响是正向的,主力合约的结果显示影响很大,月连续合约的结果显示影响比较小。持仓量对收益率波动性的影响时负向的,但这种影响比较微小,持续期较短。交易量和持仓量之间的影响较为强烈,而波动性对持仓量和交易量几乎没有影响。最后,本文使用样本外43个数据进行预测,并进行预测评价,预测评价结果显示,VAR模型比GARCH模型具有更好的效果。
[Abstract]:In order to study the relationship among yield, position and trading volume of stock index futures market, this paper studies the sample data of Shanghai and Shenzhen 300 futures contracts, main contracts and monthly continuous contracts from April 16th 2010 to March 12th 2013. The first 660 samples were used to evaluate the model and the last 43 samples were used to evaluate the model. First of all, we use GARCH model to study the effect of position and trading volume on volatility. The results show that for the main contracts, the lag term of trading volume has a significant impact on the return volatility. And the relationship between the lag term of trading volume and the volatility of return is positive, and the influence of the lag term of position on the volatility of return is also very significant. And the relationship between the lag term of position and the volatility of yield is negative. This conclusion is similar to the mainstream research. For the non-main contracts, that is, the samples of the monthly continuous contracts we selected, the GARCH model regression results show that, The relationship between trading volume and yield volatility is not significant. Then, the Granger causality test is used to test the causality between the three. The result of Granger causality test shows that there is a two-way Granger causality between trading volume and volatility. Position is the Granger cause of volatility, volatility is not Granger cause of position. There is a two-way Granger causality between trading volume and position. In order to consider the interaction of the three variables, the VAR model is established and the overall relationship of the three variables is considered. The empirical results of the main contract VAR model show that, We can still find that the lag term of trading volume has a positive effect on volatility, and the lag term of position has a negative effect on the volatility of yield. There is a strong interaction between position and trading volume, but the lag volatility has no effect on position. From the results of VAR model of monthly continuous contract, we find that volatility is mainly affected by itself. The influence of position size and trading volume is weak, while the interaction between position and trading volume is still very strong, and the lag term of volatility also has a significant impact on trading volume. Impulse response analysis shows that the main contract and non-main contract have similar conclusions, volatility has a strong impact on itself, but this effect is very short-lived. The effect of trading volume on yield volatility is positive. The results of the main contracts show a great impact, and the results of the monthly successive contracts show a relatively small impact. The impact of positions on yield volatility is negative, but this effect is relatively small. The influence between trading volume and position is stronger, while volatility has little effect on position and trading volume. Finally, 43 data samples are used to predict and evaluate the prediction. The prediction results show that the VAR model is more effective than the GARCH model.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F724.5

【参考文献】