Abstract: The absolute return of the sample with zero mean can be regarded as an unbiased estimator of volatility. Therefore, this paper chooses high frequency data of 20 consecutive trading days to study the dynamic relationship between the absolute return and trading volume of theCSI300 stock index futures.After checking the stability of the sample, the paper uses Granger test to test the causality between the intraday absolute return rate and the volume of stock index futures. The test results show that the volume of all samples is not the Granger cause of the absolute return rate, while the absolute return rate of 15 samples of 20 samples is the Granger cause of the volume.Then we use the vector autoregressive model to estimate the coefficients of 15 models, and analyze the impulse response of the estimated results. It is found that under this high-frequency sampling condition, unlike previous studies, the change of the absolute yield of theCSI300 stock index futures will significantly reduce the volume of trading during the observation period. This paper gives a reasonable explanation for the results.Therefore, in the study of intra-day trading volume and absolute yield of stock index futures, both historical trading volume and yield have an important impact on current trading volume, while absolute yield is only affected by historical absolute yield.

Key words: High-frequency data, VAR model, Grainger causality test, Impulse responses, TheCSI300stock index futures

摘要： 均值为零的观测样本绝对收益率可以看作波动率的无偏估计量,因此本文选取连续20个交易日的日内高频数据来研究沪深300股指期货日内绝对收益率和成交量之间的动态关系。在检验了样本稳定性后,文章采用格兰杰检验方法检验了股指期货日内绝对收益率和成交量之间的因果关系,检验结果表明:所有样本的成交量均不是绝对收益率的格兰杰原因,而20个样本中的15个样本的绝对收益率是成交量的格兰杰原因。随后我们使用向量自回归模型分别对15个模型的系数做估计,并对模型估计结果做了脉冲响应分析。研究发现,在此高频抽样条件下,与以往研究不同,沪深300股指期货绝对收益率的变化会明显降低其观测期的成交量。因此,在股指期货日内成交量与绝对收益率的研究中,历史成交量与收益率都对当前成交量有着重要影响;而绝对收益率只受历史绝对收益率的影响。

关键词: 高频数据, VAR模型, 格兰杰因果关系, 脉冲响应函数, 沪深300股指期货