Seasonality Effect Research Of In China's Stock Market-An Asymmetric GARCH-in-Mean Approach

Seasonality effect is the abnormal returns connected with the calendar. The existence of the seasonality is a strong challenge to the Efficient Market Hypothesis (EMH). For that reason, among others, seasonality effect has been drawing enough attentions from both policy makers and market investors.In finance literature, seasonal effect includes monthly effect, day of week effect and holiday effect, etc. Among them, monthly effect and weekly effect are studied in this thesis. Monthly effect is that the average returns in special calendar month are different from average monthly returns. As this special month is always January in lots of stock markets, we also call it January effect. Sometimes the special month is October, and it is called Mark Twin Effect. Similarly, day of week effect is that the mean returns of special days of a week are different from normal returns. Because the special day is always weekend, just like the relationship between monthly effect and January effect, weekly effect is sometimes called weekend effect.The object of this thesis for a master's degree is to study the existence of seasonality effect in Shanghai and Shenzhen A-share market. We use the return data of A-share indices ranging from July 21st, 1997 to the end of year 2000 to study this effect by employing five different asymmetric GARCH-M models. Before the GARCH analysis this paper studied the detail in very detail and find that the data is not much different from the index returns from developed market: it is fat tailed, with high kurtosis. The Ljung-Box Q test and augmented Dick-Fuller test indicate that the return series is autocorrelated and stationary. The White test accepts the hypothesis that the return serious is heteroskedasticity distributed and the Engle (1982) ARCH test indicates that both return serious has ARCH property.In the following empirical analysis, this study uses three different model specifications for GARCH-M, EGARCH-M, TGARCH-M, SAGARCH-M and GJR-GARCH-M. The paper thoroughly investigated the relation between risk, return,trade volume and seasonality effect.This study finds the following results:1. There is strong GARCH effect in both Shenzhen and Shanghai stock markets;2. Market return is positively correlated with risk. The higher the risk, the higher the return needed.3. The GARCH effect is asymmetric. The effect of bad news is stronger than that of good news4. The investors are mostly risk averse5. There is strong day of week effect in both markets. There is positive abnormal return on Tuesday and Friday, but negative abnormal return on Monday and Thursday6. There is weak monthly effect. There is weak January effect in both market and the joint significance test has reached the 10% level.