A Study About Industry Sector’s Month Effect In Chinese Stock Market

Based on the efficient market hypothesis (EMH), stock prices should follow the process of random walk in the traditional financial theory. However, more and more domestic and international empirical studies have found that an important vision called calendar effects exists in the stock market. Some developed countries gradually turn to study it, such as the United States, South Korea, Malaysia and Singapore, etc. The study index ranges from the overall index to segmentation index. While, the study about the industry sector’s month effect is few. Then, combining our countries’conditions we study the important ten industry sectors’month effect of the Shanghai and Shenzhen indices, to find out the special effects and the causes of the month effect in Chinese stock market. The special effects and the causes of the month effect play a important part in guiding investors to control their investment risk. From The article divided two plates to study the month effect:Firstly, we use GARCH models estimated by the virtual least squares method to analyze the monthly average yield of every industry sector. By comparing the accuracy and goodness of the models different GARCH models with different residual distributions the optimal model is determined. Finally, we found the EGARCH model is the best the residuals of which follow GED. Therefore, we analyze the average monthly yield of the ten industry sectors with EGARCH-GED model. Finally, we find that the monthly yield performance in different industries was different. The result relates to our countries’conditions and different industries’characteristics.Secondly, from the perspective of VaR we analyze the month effect. As far, there are few studies evaluating the month effect from the risk value. The phenomenon of leptokurtosis, fat-tail, bias, clustering and leverage exists in financial time series. Then the quantile regression approach is applied to estimate the VaR in this paper. We compute VaR model with the quantile regression for each industry sector. We select real estate sector as the representative industry to select the best model. First, we calculate the VaR under five GARCH models with three different residual distributions. The back-test is used to test the effect of VaR estimations. By comparing the results of different assumptions, we find that the proposed quantile regression VaR model is insensitive to the residual distribution of the model. Over all the situations the EGARCH model with GED distribution is the best with the lowest failure rates. Therefore,we adopt EGARCH quantile regression with GED distribution to estimate the ten industry sectors’VaR.Finally, the estimated average monthly stock returns and risk value are used to assess month effect. The study find that the month effect exists in every sector, and the features in every industry sector are different. Further more, we do some correlation analysis. Finally, some reasonable explanations are made for this market anomaly.