An Applied Study On Risk Management Of The Open-end Fund Portfolio Based On Copula Methods

The traditional correlation coefficient matrix can not be a good description of thenonlinear dependence structure of financial assets, This article combines dynamicCopula function and GARCH model on Copula-GARCH model for open-end fundportfolio risk and dependency structure. This model can effectively circumvent thetraditional linear dependence structure assumed model set error, but also take intoaccount the dependency of the dynamic changes in dependencies between assets inorder to better characterization of the structure of time-varying characteristics.Firstly, the concept and connotation of the open-end fund described to compare theexisting measure of the risk of financial market returns. Followed by application of theAutocorrelation test methods test time series of open-end fund net income to distinguishwhether the time series of open-end fund returns existence of higher order serialcorrelation. ARCH-LM test of the time series and applied to test and determine whethera higher order ARCH effect, and as a basis to select the appropriate GARCH model todescribe the marginal distribution of the open-end fund return. The distribution assumedin order to express the time series of open-end fund return distribution scharacteristics of a good characterization of the application of the KS test and AD teston the edge of the return series of open-end fund distribution model’s goodness of fittest. Again, the use of different time-varying Copula function to describe thedependency relationship between the open-end fund assets. Finally, combining theMonte Carlo simulation technique and time-varying Copula function to measureopen-end fund portfolio risk VaR, the probability of potential losses to reflect theopen-end fund portfolio risk changes to their assets in different risk situations. Weselected the three representative open-end funds as a research object, and foundtime-varying Clayton Copula-AR(p)-GARCH model estimates the VaR is more accurateand more comprehensive coverage maximum loss of the risk in each confidence level,so as to achieve the purpose of risk control.The innovation of this article is mainly reflected in: the constructed model caneffectively avoid the error of the model set by the financial time series are common fattail and partial characteristics; goodness of fit test to determine the type of optimalCopulato ensure the robustness and accuracy of the empirical results.