Font Size: a A A

Empirical Test Of The Mdh In China's Stock Market

Posted on:2006-04-27Degree:MasterType:Thesis
Country:ChinaCandidate:B C LiuFull Text:PDF
GTID:2206360152988238Subject:Finance
Abstract/Summary:PDF Full Text Request
Traditional financial theories, such as portfolio theory, capital asset pricing model and option pricing formula, all assume that the return series of the stock market obey normal distribution and have the identical variance. But mangy researchers don't think so, their empirical studies argue that the daily return series have a leptokurtic distribution and the variance is time-varying. In order to explain the empirical results, scholars have proposed some sorts of hypothesis, one of which is the so-called "Mixture of Distribution Hypothesis" (MDH) introduced by Clark(1973). This paper is to test MDH in the stock market of our country, the data used here are the daily return series and trading volume series of the market index and five stocks of Shanghai stock market from 1993 to June of 2004, high-frequency data of one year are used too. Firstly the paper examines that whether the daily return series are not normally distributed and display heteroskedasticity, secondly identifies the relation between the return volatility and trading volume with ARCH models. Then, we try to use MDH to explain the empirical results above, and conduct explicit tests of two models of MDH. Finally, according to MDH, this paper puts forward an idea of how to get normally distributed return series, and test it with high-frequency data.The main research conclusions of this article are as follows:Firstly, the daily return series of our country's stock market are not normally distributed and display strong conditional heteroskedasticity. The distributions of return series of market index and five stocks are all positively skewed, the J-B statistics are very large, and the statistically significant GARCH coefficients are near tol.Secondly, trading volume can explain return volatility a lot. Including trading volume in the conditional variance equation does result in a reduction of volatility persistence for four stocks, and the coefficient of trading volume is positive and statistically significant.Thirdly, in our country's stock market, we reject the two explicit models of MDH, but we can not come to the conclusion of rejecting MDH.Lastly, according to MDH, it is possible to get normally distributed return series.
Keywords/Search Tags:return, trading volume, volatility, MDH, normal distribution, heteroskedasticity, ARCH, GARCH, GMM
PDF Full Text Request
Related items