Font Size: a A A

Long Memory And Non-Liner Time Varying Correlation Between Price Volatility And Trading Volume In Chiniese Stock Market

Posted on:2013-11-05Degree:MasterType:Thesis
Country:ChinaCandidate:Y L HuFull Text:PDF
GTID:2249330377956659Subject:Applied Mathematics
Abstract/Summary:PDF Full Text Request
The relationship between trading volume and price in capital market laid technical analysis as core position in the modern securities investment.Mixture Distribution Hypothesis provided the theoretical basis for the relationship between the volume and price. Compared with practical analysis,theoretical analysis results between the volume and price could be improved.Basing on the long memory fractal thought,multivariate GARCH fluctuation model and financial temporal nonlinear time-varying related Copula theory,we studied the further relationship between the volume and price. the main work was as follow:On the one hand, the existing research mainly took the volume as exogenous variables of return fluctuation equation to study the linear relationship between the volume and price. Trading volume was seen as a main subject as well as return rate and we also considered the volatility persistence. Using multivariate GARCH and Copula theory research methods,we further and systematacially studied the relationship between the volume and price.Using the method of long memory financial time series,we made long memory test for the Shanghai index logarithm return and trading series. The results supported that two level sequence didn’t exist long memory, but fluctuation sequences represented strong sustainability.We built the fluctuation models of long memory for logarithm returns and trading series.Then we compared them with short memory models. Models results showed that the fluctuations sequences of logarithm returns and trading series both represented long memory,this suggested that short memory models were insufficient to describe temporal fluctuations.At the same time there were similar fractal parameters for two series,that suggested that they had the same persistence.Next,using multivariate GARCH model thought,by applying the bivariate GARCH model to analyze the relation between trading volume and price in the Chinese stock market,the models results showed that the BEKK models was optimal.It suggested there was significant positive correlation between trading volume and price.On the other hand,using copula theory that measured nonlinear、tail and time varying correlation,we analysed the relation between trading volume and price in the Chinese stock market and proposed a SJC TVP Copula-FIGARCH-t model which could measure two timing series heteroscedastic,long memory,nonlinear time-varying, tail correlation. The empirical analysis results on the Shanghai index logarithmic rate of return and trading volumes series showed that our model was superior to other existing models. The results showed that the Shanghai index in the sample period the trading volume and logarithm return them self were more close to the t distribution rather than normal distribution,there were long memory and long memory parameters were approximately equal.There were nonlinear upper tail correlation and obvious time-varying positive correlation between price and volume.
Keywords/Search Tags:Long Memory, FIGARCH, Multivariate GARCH Model, Copula Theory, SJC TVP Copula-FIGARCH-t model
PDF Full Text Request
Related items