The Empirical Research On Calendar Effects Of Security Returns In China

Efficient market hypothesis (EMH) is the cornerstone of modern financial theory, which was deepen and made by Fama (1970). He defined efficient market that in a stock market, which prices reflect all available information, then we call this a market for the efficient market. But more and more vision on the financial markets began to challenge the efficient market hypothesis. In this financial context, the vision of the market research studies over the past few years become an important part of market efficiency.An important vision of stock market is the calendar effect. Calendar effects was found in the study of the initial rate of return on equity, which mean that a specific time period rate of return significantly above or below average return. People called this vision, because they can not be explained by the traditional asset pricing models. This type of effects including month effect, quarter effect, week effect, holiday effect, and so on.This sample data is the Shanghai Composite Index closing price from January 1, 1997 to December 31,2010. We study four kinds of China's Shanghai stock market calendar effect, they are:month effect, monthly rotation effect, week effect and holiday effect. We studied both the overall sample calendar effect and the rolling sample calendar effects. We use rolling sample test method, which in addition to the 14-year intervals to test the overall sample, but also divided the sample into 10 rolling samples, Overall sample (14 years) trying to explain the persistence of China's stock market calendar effects, and sub-samples (5 years) calendar effect attempts to explain the changes of calendar effect in our country. Dummy variable method is used in the empirical method.We obvious found in the empirical test of China's stock market calendar effect on the entire sample period were:positive "March effect"; positive "the beginning of month effect"; positive "Wednesday effect" and negative "week four effect" positive "Labor Day" effect and the "Spring Festival effect". We used the method of rolling sample tests, the results from the regression coefficient of Z statistics charts show the presence of many other effects. In the sub-sample study of calendar effects, we found that some calendar effects are not stable, they will change over time, become strong, weakened, disappeared, or even reverse.We explained these calendar effects from the policies and systems, consumer behavior, cognitive bias, arbitrage-driven, and cash-rich investors. Besides, this paper proposes an investment strategy:Holding stock in March, the first five days a month, three days before and after the Spring Festival, Labor Day, National Day, you can get excess returns. For non-validity of this market, we made policy recommendations:improve the efficiency of stock market information disclosure, standards of information disclosure system, strengthen the Government's policy of information dissemination management, regulate the behavior of intermediary service organizations, enhance investor training, and strengthen the normative management of listed companies.