The Jump Behavior Of China's Stock Market Prices

The behavior of stock price is the basis of asset pricing,portfolio choice,risk management and product designing.In 1973,based on the assumption that stock price follows geometric Brownian motion(GBM),Black,Scholes and Merton deduced the famous Black-Scholes option pricing formula.But later researches found that many stylized facts of asset return, such as leptokurtosis and fat tail,can not be captured by GBM;moreover,the Black-Scholes model is inconsistent with the real data.Therefore,researchers introduced another stochastic source,jump,which is totally different from Brownian motion.Empirical studies confirm that the discontinuities of stock price movement,jumps,are essential in financial market, especially in emerging markets.But,few researches focus on this topic in China.This paper is on the jump behavior of Chinese stock price,and aims to fill the gap in the study of Chinese stock market.Using the 5-minute high frequency data of the Shanghai Stock Exchange Composite Index(SSECI) for the period of March 2001 to February 2008, our subject consists of two parts.The first one is parametric model based on Ma(1992),we derive the moments condition and estimate the parameters of the geometric Levy process, which is the most common jump-diffusion model,and the impacts of jumps on the return distributions are analyzed using Monte-Carlo simulation and inverse Laplace transformation. The second part follows the non-parametric technique proposed by Barndorff-Nielsen and Shephard(2004b,2006a),we investigate and carry out various statistical tests on the existence of jumps for the SSECI.The location(time),size distribution of the jump component separated from the high frequency data are analyzed along with discussions on the causes of several recorded extreme jump events in the sample period.Our findings can be summarized as following:1.There are frequent and significantly rising and falling in Chinese stock market,and the leptokurtosis characteristic of the returns distribution is apparent.2.Jump process can well capture the leptokurtosis characteristics of the stock returns,and existence of the jump component in the Chinese stock market is thus confirmed;furthermore,the parameters estimation of the geometric L(?)vy process shows there is a considerable amount of jumps in Chinese stock market.3.The non-parametric model analysis shows that daily jumps constitute 7-17%portion of our sample,and contribute about 30%of the total variance;the distribution of jumps is asymmetry in time and size scales,and displays jump clustering;comparing to the bear market, more jumps take place in bull market,and most of the times move upward.4.The 'jump' adjusted return series obtained by separating the jumps from the original time series was modeled under a parametric specification of the geometric LeVy process.Significant statistical evidences were found in supporting the hypothesis that the adjusted return series are nearly normally distributed.5.Although some of the jumps are triggered by financial events,the causes of jumps are,in general,complex and obscure.