The Study Of Portfolio Selection And Risk Forcasting For Chinese Financial Markets Based On Vine Copula Theory

Economic globalization and capital internationalization makes the relationship between different economic subjects more and more complex. Investment globalization and uninterrupted trading mechanism strengthened the linkage among the worldwide financial markets. The domestic financial markets in China have got vigorous development along with the reform and opening up and have already taken shape the financial market system with the characteristics of multi-level trading places, diversiform trading categories and multiple trading mechanisms. Many financial products are publicly traded in the market, such as stocks, bonds, funds and commodities different financial products and variety transaction mechanisms exist at the same time, such as spot, future, repurchase and so on. The differences of products and trading mechanisms make the relationship between different financial markets more and more complexity, which clearly put forward higher requirements for investment decision-making and financial regulatory policy-making. Therefore, there is important theoretical and realistic significance to carry out the further research on portfolio selection, risk prediction and related issues on the basis of the relationship study among the different financial markets of China from a comprehensive perspective with multiple markets as the research object.This paper took the bond market, the stock market, the fund market, the stock index future market, the foreign exchange market, the money market, the commodity future market and the spot gold market into a unified framework and with the daily closed prices of bond index, HS300 stock index, The Shanghai stock exchange fund index, HS300 stock future index, the middle contract price of euro against RMB index, weekly interbank lending rate index in Shanghai, Shanghai copper future index and the purity of 99.95%gold spot from April 16,2010 to November 30,2014 as sample. First of all, the paper empirically analyzed the causality and the dynamic relationship between different markets, and then estimated their marginal distribution with GJR and EVT theory, and modeled the connect structure among each financial market using C-Vine, D-Vine and R-Vine copula, and detail analyzed the net relationship of every two markets under the condition of any other one market, last predicted the risk of portfolio.VaR and ES. based on the results of the connect estimation of the Vine copula with the Monte Carlo simulation method. The empirical analysis draws the following conclusion:(1) The causal relationship between different financial markets in China is rare, which indicts that the market system lacks corresponding logic connection, so the government should further strengthen the construction of relevant market to optimize the structure of the financial system, increase the investment and financing options and provide more channels for market regulation. (2) The correlation between two different financial markets in China are different in features and the size of fluctuations, so building dual portfolio in bear period with the markets whose correlation are stability and performance lower (or negative) could circumvent the two market prices fall at the same time. (3) In addition to HS300 index and Shanghai stock exchange fund index, HS300 stock index and HS300 stock future index, foreign exchange index and Shanghai copper future index, Shanghai copper future index and gold spot index four portfolios, the unconditional correlation of other every two markets are greatly reduced in considering the third conditional market, so building the ternary portfolio with the corresponding market could achieve the purpose of avoiding risk effectively; The conditional correlation between every two financial markets under the condition of multiple markets (especially when the conditional markets more than two) all less than their unconditional correlation, so building a diversified portfolio with these conditional markets rather than invest in the two markets direct with higher unconditional correlation when considering long or short at the same time in these two markets could avoid the risk of market prices fall at the same time. (4) C-Vine, D-Vine and R-Vine copula connect models have no significant difference in fitting the dependency structure of Chinese financial markets, and the VaR of Chinese financial markets portfolio at the quantiles of 0.95,0.975, 0.99 and 0.995 forecasting combined with the results of Vine structure estimation and Monte Carlo model are more accurate than conventional models of H-S, M-V and DCC. (5) Only C-Vine and D-Vine copula in three Vine copula connect models can accurately predict the ES of Chinese financial markets further on the basis of VaR forecasting, and D-Vine copula connect model under a relatively high quantile levels has more precision. Robustness analysis demonstrate that the empirical results for portfolio selection and risk prediction of financial markets in China based on the Vine copula connect models are reliable, and the accordingly policy suggestions could solve the practical problems.It has a certain instruction value that portfolio selection, risk prediction, fund management evaluation and other related laws and regulations formulation with this article research conclusion on Chinese financial markets, and the research method has significance for the related work of Chinese financial market portfolio risk management. But the problems of Chinese financial markets marginal distribution modeling, actual Vine copula connect model building with considering the characteristics of the financial system in China and the portfolio optimization on the basis of Vine structure are yet to be further in-depth study.