| In order to provide investors with a means to hedge against risks when investing in the security market,China Financial Futures Exchange(CFFEX)launched a mock trading on Oct.30,2006 for CSI 300 Stock Index Futures,which were listed and traded on Apr.16,2010.CSI 300 Index Futures Contract was the first financial future product formally launched in China and it opened a new era for China's securities market.So a comprehensive,dynamic research on our country's stock index and stock index futures has practical significance and theoretical value.From the Angle of time-varying,With the theme of dynamic dependence between stock index futures and spot and dynamic hedging ratio,It has carried on the deep discussion and empirical analysis on the futures and spot dynamic dependency structure,the dependence degree of measurement,the futures and spot earnings volatility measurement and prediction,the volatility of futures and spot and basis,volatility and dynamic effect of the relation between time-varying correlations,and dynamic hedging ratio.The aim is trying to depict the dynamic effect of relationship between stock index futures and spot,and explore the optimal hedging ratio and provide reference for marketing managers and investors,and for our country has been listed in the national debt futures and many related research will be listed on the interest rate futures.The first part gives the basic statistical description of two time series sample data of stock index futures and spot,Including normality test and stationarity test for the price series and return series of the two financial assets,cointegration test between the two price series,identification of ARMA models of two return series and the ARCH effect test of residual.The results show that,the price series of the stock index futures and spot are stationary series,and there is a cointegration relationship between the two price series;Both return series are weak stationarity,and have statistical characters of obvious peaks,fat tails,non-normality;Stock index futures and spot returns are subject to average model of ARMA(0,0),and the residual of models have ARCH effect.The basic statistical analysis characterized the features of stock index futures and spot.And it indicates that the traditional correlation method is no longer applicable,and it will provide the necessary reference for the establishment of the statistical model in the following sections.The second part introduces the copula models which can depict the nonnormality,nonlinear dependence structure,and nine copula models are built for the joint distribution of the stock index futures and spot,it found that t-copula,SJC-copula,Rotated Gumbel copula and Normal copula is superior to the other five constant parameter copula model.By convert the four model constant parameters to time-varying parameters,four time-varying copulas connect model is established,The empirical results show that,the t-copula model is the relatively best of all,and dynamic dependence of stock index futures and spot is measured by time-varying t-copula model,and evolution equation was gained,the relationship between the price basis and the correlation were analysed.The research shows that a dynamic dependence exists between the first financial futures market and spot market in China,which has the characteristics of persistence and fat tail to a certain extent,but since the ability to adjust to each other is existing between the financial markets,the dependence will remain around the mean value rather than enhancing or weakening persistently.In addition,besides the impact of the financial asset returns,the dynamic dependence between stock index futures and spot has a lot to do with the fluctuation of the basis between the two asset prices.Especially when a large fluctuation occurs in the spot return,or even in the price basis,the dependence between the two markets will be weakened sharply,and the hedge risk effect for the financial futures will be reduced correspondingly.This part of the established dependence model provides the basis for the subsequent dynamic hedging model selection and establishment.in order to more accurately describe volatility of stock index futures and spot,the third part compared the GARCH,garch-m,EGARCH three model of different characteristics with the hypothesis of three kinds of residual distribution respectively,and the volatility of each point time were measured.Selection from June 2014 to December 2014 as forecast period,in-sample forecasts,steps forward out-sample forecasts,one step forward out-sample forecasts for the volatility of stock index futures and spot returns,and realized volatility based on high frequency data of forecast period were given.Forecast information of the three forecasting volatility was explored through single variable,double and three variables regression model of realized volatility and forecasting volatility.The results show that the model with residual of biased students t distribution is superior to the other models,leverage effect was not supported by empirical results,and the relationship between the expected return and volatility was not significant,and its reaction is slow to the impact of the external effect,the volatility of stock index futures has certain feedback effect,the greater the volatility,the bigger the expectation,and the influence of the volatility effect has strong sustainability,drastic fluctuations in the market will be difficult to immediately eliminate in a short time;one step forward rolling forecasts contains more forecast information than the other two.This part of the measurement of volatility will also be preparing for the following dynamic influence factors analysis and optimal dynamic hedging ratio research.The fourth part,based on the second part of time-varying correlation coefficient and the third part of the measure of volatility,by using TVP-VAR models,obtains the impulse response of the futures and spot return volatility and basis,and that of earnings volatility and time-varying correlation coefficient,and analyzes the dynamic interaction relations of volatility,basis,and time-varying correlation.Empirical shows that,after 2011,there is mutual influence between the stock index futures and spot two return volatility,and the impact of stock index futures returns volatility impact is far less than the impact of the volatility of the spot;The influence of the impact of basis on the volatility of stock index futures and spot is close to the external economic environment,and when market is not so optimistic,basis fluctuations will lead to increased volatility,and the market is optimistic,basis of fluctuation will reduce volatility of stock index returns;When the correlation is enhanced between futures and spot markets,stock index and stock index futures return volatility will weaken,the market will be more stable.The fifth part calculats the optimal hedging ratio based on the traditional linear regression model,classic bivariate GARCH and time-varying copula-GARCH model,and three kinds of hedging portfolio are constructed through the hedging ratio of each model,comprehensive evaluation of the different hedging model with different hedging portfolio is given.The results indicate that the dynamic hedging model is better than traditional static model in terms of the effect of hedging;From the perspective of minimum variance of edging portfolio income,the ECM-DCC-GARCH model is best;From the perspective of maximum average return of hedging portfolio,time-varying copula-ECM-GARCH model is best;current hedging portfolio hedging is superior to the rest of the two kinds of hysteresis.Main contribution and innovation of this article is the study of dynamic correlation from the view of time-varying,and characteristics of the dynamic correlation evolution equations are obtained;TVP-VAR model is a leading use for analyzes the dynamic interaction relationship between volatility and basis,as for volatility and time-varying correlation;the time-varying copula-ECM-GARCH model is built for dynamic hedging ratio that can gain more return. |
PDF Full Text Request