| Any financial product investments have risks, and exchange rate risk is double-edged sword, the risk from uncertainty, both may be damaged so that enterprises may also benefit from it. More and more retail investors and enterprises began to enter the foreign exchange market, you can take advantage of fluctuations in the foreign exchange market, exchange rate fluctuations can be used to earn money. Especially for export enterprises to reduce the foreign exchange risk, while working closely with the relevant government departments, banks, credit insurance departments to maintain communication, timely informed of warnings about the risks and improve the ability to guard against foreign exchange risk. So its risk analysis and estimates highlight the important measure. Correlation analysis is a multivariate analysis of a central financial issues, asset pricing, portfolio, fluctuating conduction and overflow, risk management and other issues related to correlation analysis. But the traditional linear correlation coefficient has certain limitations. For the analysis of the financial asset portfolio, its relevance and its joint use of an effective distribution model is particularly important..Aspects of risk measurement, VaR has become the most widely used method, but also financial risks focus of the study; through a portfolio of financial assets Copula build joint distribution function to create a convenient, scientific method, so that the joint distribution of financial assets to meet the inherent fat tail characteristics-non-normal, non-linear correlation can solve the traditional risk management model is a linear correlation between normality assumption.This paper is primarily aimed at multi-financial assets portfolio correlation analysis, measurement problems, an example to study Copula theory and application of modeling methods. Based on Copula theory, through the Copula function, GARCH models, VaR and Monte Carlo methods combine to solve the multi-financial assets, non-normality non-linear correlation modeling, and through the use of nested Archimedean Copula establishment of high-dimensional portfolio model. The first part of the first empirical study object is the Chinese foreign exchange market foreign exchange assets of several major investment portfolio, first through GARCH Models comparative study to determine the risk of a single marginal distributions of capital gains rate volatility model; then use the PC algorithm to estimate the correlation between assets expressed structure, based on the nested nested Archimedean Copula modeling idea to build a high-dimensional nested nested Archimedean Copula model, which can better describe the dependencies between the portfolio structure; nested embedded in high dimensional sets Archimedean Copula model based on Monte Carlo simulations using the portfolio VaR, and through back testing proved the validity of the model. Followed by the second part of the empirical study is based on the four kinds of Chinese foreign exchange market portfolio as an object, in this part of the study based on high-dimensional modeling methods Pair Copula mixed C&D, rattan cane Copula model Copula model comparative study, empirical exchange VaR of the portfolio assets. In the empirical study, the two types of models in the portfolio VaR calculation accuracy for comparison. Two-part empirical results show that the established Copula model can better characterize the correlation between the structure of financial assets, in order to provide convenient conditions for risk measurement. |
