论文部分内容阅读
基于分位数回归(QR)模型分析了不同分位水平下的收益率与成交量的关系,考虑到收益水平对成交量的影响,引入工具变量,构建IVQR模型,更加客观地分析了不同分位水平下成交量对收益率的作用。蒙特卡洛模拟结果表明IVQR估计比QR估计具有更小的偏差和更强的稳健性。实证分析结果表明条件收益率处于较高水平时与成交量正相关,且分位水平越高两者之间的相关性越大;条件收益率处于较低水平时与成交量负相关,且分位水平越低相关性越大;多数分位水平下成交量对收益率的影响并没有较大的差异。同时结果还表明工具变量的分位数回归模型能较好地处理模型中的内生性问题。
Based on the quantile regression (QR) model, the relationship between the yield and trading volume at different quantile levels was analyzed. Taking into account the influence of income level on the trading volume, a tool variable was introduced to construct the IVQR model, which further objectively analyzed the different points The level of volume under the impact of yield. The Monte Carlo simulation results show that IVQR estimation has less deviation and stronger robustness than QR estimation. The results of empirical analysis show that the conditional return at a high level is positively correlated with the volume, and the higher the level of the quantile is, the greater the correlation between the two is. The lower the conditional return is, the lower the volume is and the volume is negatively correlated The lower the level is, the higher the correlation is. There is no significant difference in the effect of volume on the yield at most sub-quantile levels. At the same time the results also show that the quantile regression model of the tool variables can better deal with the endogenous problems in the model.