High-frequency Data Based Empirical Study On Price Discreteness In China Stock Market

With the establishment and developing of China's financial database, we can get more high frequency financial transaction data at lower cost. The frequency less than a day's trading data, such as each trading data, five minutes trading data,fifteen minutes trading data and so on, the data is high frequency data. The available high frequency data make the study on market microstructure further.Price discreteness is the important characteristics of financial market microstructure. It is not an important question to low frequency data, because it can be used as a continuous process approximatively. The traditional financial market microstructure study followed the study method,the methods will be lost many market information undoubtedly, resulting in market research appear bigger deviation. In fact,to high frequency transaction data sample,price discreteness is a serious problem, because the observation value of price movements is limited, so in continuity assumption the model also no longer apply. This paper, by using each trading data and five minutes trading data, analyzed share price discreteness from theory and experiment angles, and used the ordered probit model to discrete share price model, analyzed the factor of Stock intraday price changes.Part 1 of this article is an introduction. This part briefly introduces the research background, research significance and literature review. Part 2 mainly analyses the influence of discreteness to price clustering and stock yields, selectes two high-priced stocks and two low-priced stocks, a mid-priced stocks, using five minutes trading data in September 2010 and the sub-transaction data in August 2010 analyzes the price clustering phenomenon, the distribution features of price changes and 5 minutes yields two-dimensional historical relationship diagram, obtaines the conclusion that the low-price stocks more susceptible to the price discreteness than the high-price stocks. Part 3 briefly makes a comparative analysis for several discrete model, it obtains that the ordered probit is the optimal model for capture the price discreteness,and finally introduces the basic idea of ordered probit model which proposed on the basis of the discreteness. Part 4 is the empirical analysis of trading data. This part mainly describes the selection and grouping of the data, the stock intraday price changes model's design, empirical results and empirical analysis. Part 5 is the conclusion.This part mainly does a simple sum on the analysis of the preceding sections, and gives the relevant policy Suggestions.The main contribution of this paper to the study of market microstructure in our country is embodied in the following aspects:Firstly. This article constructs discrete model based on the high frequency trading data of China's stock market for empirical analysis, and concludes that the influence of discreteness to low-priced stocks is larger than high-priced stocks, information cost and inventory cost have an important impact on intraday price movement, and in China's stock market the impact of information cost is greater than the inventory cost.Secondly, This article's data selected is each trading data in August to October, 2010. In recent years, although the home also has some scholars to do empirical analysis about price discreteness, the data were 2006 years ago. In fact, since 2006, China's stock market has undergone great changes, such as the fluctuations become larger, more frequent transactions (most of the daily trading data of shares up to 4000 times more) and other features. Obviously useing trading data of 2006 years ago on empirical analysis was lack of timeliness, the conclusion are also much different.Thirdly, The selection of variables in this paper, in the use of foreign research methods combined with the institutional background of China's stock market, introduces four explanatory variables such as information cost, inventory cost, liquidity cost and accumulative pressure of buy or sell in ordered probit model.Fourthly, In this paper, specific to the result of empirical analysis on stock price discreteness, puts forward relevant policy suggestions and bidding strategies to the securities regulatory authorities and investors.Finally, This article is earlier summarizes the development of the stock price discreteness theory and empirical research, especially points out the important influence of information cost on price movement,reflects that China's stock market commonly exists the characteristics of information trading.