The Stock Market "week Effect" Statistical Research

Many empirical investigations have validated that the "day-of-week effect" is an ubiquitous phenomena in many developed countries' securities business and in some burgeoning stock market. But the methods investigations have used were so different that one conclusion can not be horizontally compared with the other conclusion. Furthermore, some method model themselves were possible ill-suited to securities data. How to establish a more reasonable statistical testing method is the issues what this dissertation carefully investigates to solve. The original ideas of this dissertation are the following:(l)Based on the investigation of the existed testing method for the "day-of-week effect" in literature, we put forward a new comparatively rational testing model in theory. And, we empirically investigated the "day-of-week effect" on Chinese stock market with using the practical data of Shanghai and Shenzhen stock market. The conclusion is that Chienese stock market exists the different formal the "day-of-week effect" in divided stages.(2) When analyzing the testing method for the "day-of-week effect", we considered the OLS regression effect and the testing power will decrease becauseof the different variable in real securities data. We considered the identification.estimation and hypothesis testing of the linear regression model with one rankauto-regression (AR(1)) error(3) We considered how investors reasonably utilize the testing result of the"day-of-week effect" to direct the investment strategies after obtained it.Connecting with moving average line, which is usually used by investors, we established three strategy models. Computer program simulating results haveshown that the three constructed strategy action would provide investors'decision-making with assistance information, and among those, the third strategyaction are the optimum strategy.(4)When we considered how to provide the practical investors for directionswith using the testing result of the "day-of-week effect", we connected the indexof moving average line. We summarized the characteristic and application of moving average line in the stock market. And this dissertation puts forward a new ameliorated model with moving average line connecting the practice of stock market. The result of computer program show that using our model to invest can validly decrease risk and increase return.At the same time, we have solved the practical problem of the Cox-Stuart trend testing method, which can supply a direction for using this method in practice.