Font Size: a A A

Study On Risk Connection Of Multimarket Base On Semi-parametric Dynamic Vine Copula

Posted on:2018-02-13Degree:MasterType:Thesis
Country:ChinaCandidate:J B LiangFull Text:PDF
GTID:2439330512986064Subject:Financial engineering
Abstract/Summary:PDF Full Text Request
Based on the idea of Hafner and Reznikova,this article extends the semi-parametric method to vine copula function and construct a SVR-dynamic vine copula system combined with support vector regression,which can analyse the dynamic dependent structure between high-dimensional variable.Regarding the volatility index as a measurement of risk,this article analyse the indexes of Hong Kong,Korea and America from Jan.1st,2004 to Aug.31st 2015 and test whether the model could accurately identify the dynamic dependent structure of high-dimensional variables.And here is the conclusion:First,the effect of risk connection amongst Hong Kong,Korea and America is asymmetrical.Using the static vine copula model on the volatility indexes,I find that the risk connection and panic mood between Hong Kong and America,Korea exist the phenomenon of up tail correlation structure,namely one market holds pessimistic expectations then another market will be pessimistic probably which mean risk contagion exists.Known expectation of Hong Kong,the market of Korea and America exists the phenomenon of low tail correlation which mean that there is a certain interaction between the America and Korea.The result of stress test also confirms the asymmetric correlation and America market is more likely to be affected.Second,the asymmetrical effect of risk connection amongst three markets is dynamic visibly because of the following aspects:1)the parameters of the copulas generally fluctuate up and down with constant parameter values and there are peaks and troughs.2)the dynamic parameter values is bigger than constant parameter values regularly.3)the parameters are in high level during crisis which mean no matter high or low volatility exception one market will be affected by anther markets.4)the structure of the dependence during the advanced financial crisis is similar.Last,from the similarity of the dependent structure between nowadays and financial crisis period,the model can reflect the anxiety about crisis.In the end of sample period there is the anxiety about crisis,the SVR-dynamic vine copula model discover the similarity of the structure which mean the anxiety is justified and the model can capture the message.
Keywords/Search Tags:VIX indexes, time varying vine copula, semi-parametric, support vector regression, panic
PDF Full Text Request
Related items