Dependence And Tail Dependence Between Shanghai And Hong Kong Stock Market Based On Copulas

With the globalization of economics and the dependence of financial markets to beenhanced，the study based on unitary asset or market is not always reaching theneed．The relevant analysis becomes more and more important in the financialapplication, especially in the area of financial risk analysis．It has become a key riskmeasure．But the former research based on liner correlation usually all concentrates inthe degree of relativity, and neglects to related structure or the tail dependence in themoney market．Granger causal analysis is qualitative，and can not give quantitativeanswers．We use the non-linear correlation analysis tool----Copula to study thedependence of financial markets and to analyze the constant dependence, time-varyingdependence and tail dependence of Shanghai stock index and Hong Kong stock index．This paper’s content and structure of the arrangement is roughly as follows:The first part of this paper introduces the background, significance and situation ofthe research; The second part introduces some basic knowledge of Copula; The thirdpart introduces modeling, parameter estimation and the inspection; The fourth partintroduces the static and time-varying Copula functions used in this paper; The fifth partwas empirical research; The six part is conclusion.In the research method, the existed research usually assumpted that the edge isnormal distribution or t distribution, under the premise of static Copula method to thedependence of the two cities or the tail dependence, generally use the parameterestimation method (Parametric Approach) to the specific Copula function for parameterestimation, parameter estimation method of distribution function in the edge to edgedistribution function of micro roughness itself was made relevant assumption, thatassumption that edge distribution function is normal distribution or t distribution and soon, but the assumption that it is difficult to accord with the actual situation, so for edgefitting effect and the distribution is not very good, this often affect the parameterestimation effect of Copula function. MLE based on ranks was used for the marginaldistribution, through the static Copula and time-varying Copula to study the dependenceand tail dependence of Shanghai and Hong Kong stock market. The results show thatthe dependence of the two stock market presents general growth trend; The resultsshowed that under the given sample the lower tail dependence is slightly larger than theupper tail dependence. Followed by analysis of the sample was divided into two phases, the first phase found almost no correlation of the lower tail, and the upper taildependence is not obvious; the second phase of the tail (especially the lower tail)associated with significantly higher than the first stage, and the lower tail dependence isstronger than the upper tail dependence obviously.