Font Size: a A A

Research On Semi-parametric C-Vine Copula Model With Measuring The Structure Of Finaical Risk

Posted on:2017-08-31Degree:DoctorType:Dissertation
Country:ChinaCandidate:Z Q ZhangFull Text:PDF
GTID:1319330515489369Subject:Quantitative Economics
Abstract/Summary:PDF Full Text Request
In 2016,as the first year of the Thirteenth Five-Year Plan,Chinese economy has entered a"New Normal"."New normal" is a process of dynamic optimization rather than a fixed status.In this process,the indicators of economy and financial market present the new features;the associations of its inherent risk get more complicated.As a result,the traditional econometric models are difficult to measure risk dependent structures effectively;new theoretical tools are required to solve the problem of measuring risk structures at "New Normal".Since semi-parametric C-Vine Copula model connect with features of robust,flexible and easy to be estimated,it has the advantage to solve the above problems,the research of this function is necessary for theoretical meaning and application background.Copula function model is a new kind of modeling tool for joint distribution of multivariate.Its remarkable feature is two step structures,making the individual characteristics of each variable is embodied in its marginal distribution,the dependent structures are embodied in Copula functions.Compared with the traditional multivariate normal and multivariate t distribution,the model relaxes the assumption,it has more flexible and convenient structure,more widely application,and more capacity to fit the joint distribution of economic variables truly.In view of the above advantages,according to the characteristics of the structure of financial risk at present,the research of Copula function model in paper as follow:At the beginning,the Copula function of the domestic and foreign literatures are sorted and summarized.Moreover,the theoretical properties and application scope of the tools used in measures of association,marginal distribution,Copula function model,estimation and test are studied in detail.Once again,the research combines nonparametric kernel density estimation method with popular Vine Copula model,constructs semi-parametric C-Vine Copula model.Finally,base on the above study,risk dependence degree and structure of stock market in five countries which are including United States,China,Japan,Germany,Britain are measured.And its economic meaning is discussed and analyzed.The paper's main innovation is reflected in the following aspects:firstly,it is proposed that the STAR model can be used to fit the marginal distribution,and then the STAR-Copula model can be constructed in the case that the sample size of data is small or inherent nonlinear transfer mechanism are presented at data.Secondly,in the part of research on nonparametric kernel density method,for the second order,higher order and asymmetric kernel function are carried on the theoretical research and simulation in various distribution environments,it is proved that the normal and asymmetric kernel have better estimation properties;biweight,triangular and Epanechnikov kernel have smaller boundary effect.Thirdly,combining the advantage of fewer data constraints for the nonparametric kernel density method and the advantage of flexible construction for the Copula Vine model,semi-parametric Copula C-Vine model is originally proposed,and then estimation properties of the model under various conditions are simulated by Monte Carlo method,it is proved that the model has good robustness,and it is suitable for the multivariate joint distribution modeling under the condition of the data distribution irregularity and data volatility unusually.The innovations of paper provide powerful quantitative analysis tools for the measure research of financial risk dependent structures under"New Normal".
Keywords/Search Tags:Vine Copula Model, Kernel Density Estimation, Semi-Parameter, Dependent Structures, Risk Measurement
PDF Full Text Request
Related items