The Measurement And Empirical Study Of Portfolio Risk In Chinese Stock Market

With the volatility of international financial markets, international investors who are in urgent need to strengthen financial risk management techniques; important factor in financial market risk has always been all investors pay close attention to, and effectively measure financial risk each investor was scientific and rational investment provides a theoretical basis; due to the unique characteristics of the financial market volatility, long memory, such as fat tail, volatility clustering of yield, volatility spillovers, leverage and volatility and persistence, etc., so that greater financial risks accurately measure the difficulty.1952established Markowitz mean-variance theory for the first time to quantify the risks and benefits, access to the calculation of risk minimization, open the door to financial risk measurement. However, the mean-variance theory there are certain financial assumptions inconsistent with the actual facts, not a good description of the financial characteristics. VaR (Value at Risk) technology is the development of value-at-risk after the1990s a new type of risk measurement methods, with a wide range of applications, and traditional risk measurement methods, compared with a higher utility value. However, there are many problems VaR technology, such as it does not take tail risk, investors are likely to make real use of higher risk investments, does not have a sub-additive contrary to the principle of risk diversification is difficult to apply to portfolio optimization. For lack of VaR technology, theorists proposed many improved methods, ES (also known as C-VaR) risk measurement method is one of them. ES risk measurement method have a valid measure of the loss of the loss of value on the VaR levels, can fully measure the tail of the loss, with sub-additive properties, can be easily adopted in the portfolio moderate advantage.Fluctuations in the financial forecasting model sequences into two main categories:GARCH family models and SV family model. Engle (1982) proposed ARCH model can effectively describe the characteristics of volatility clustering of financial time series. Bollerslev (1986) further proposed generalized ARCH model (ie GARCH), which not only can describe the characteristics of volatility clustering of time series, but also a good catch fat tail characteristics of the financial datadistribution. As the SV model does not rely on historical information, many scholars confirmed SV model is suitable for practical research introduced in the financial sector. And domestic and international research shows that there are more SV model has the ability to portray better than GARCH model financial time series. Then explore the relevance of the relationship between financial markets, the correlation between the financial markets is very important, particularly in relation to asset pricing, portfolio volatility spillover effects of conduction and other issues Correlation analysis methods commonly used linear correlation coefficient, Granger causality analysis, but they are a lot of deficiencies:Features (1) financial time series showed a simple linear relationship does not meet the financial facts, can not be analyzed variables exist between non-linear relationship between the variables and mainly for the non-linear relationship;(2) Granger causality method is generally only draw simple conclusions qualitative causal relationship can not be described quantitatively. To solve this problem we introduce Copula function, which was first proposed by Sklar (1959) proposed. Copula function as a new statistical method to analyze the correlation between variables, are widely used non-parametric statistical areas, especially in the study of the correlation between random variables is very reliable. His direct characterized edge variables as a function of a variable distribution is analyzed to select the edge of the distribution without restriction. Frees and Valdez (1998), Embrechts (1999) Mcneil and Straumann (1999), who will Copula technology into the financial sector, the financial sector for further research model provides a new idea.This paper mainly deals fluctuation characteristics of financial markets, using the GARCH family models and SV models Shanghai Composite Index and Shenzhen, as referring to the volatility of financial data to predict and value at risk VaR, expected loss measure ES effectively portray our stock market volatility main features:fat tail, volatility clustering, leverage effect, spillover effects, etc.; also confirmed that ES metrics advantage compared to VaR. By volatility model yields a sequence of edges distribution of accurate description, then Copula function simultaneous introduction of both the Shanghai and Shenzhen index returns accurately depict the distribution of their dependency structure, and then calculate the portfolio VaR and ES values and comparison confirmed that ES risk measures under Copula function of portfolio risk measure for better applicability, and won the best ratio of investment risk in the portfolio value of the minimum scenario. In this paper, under the guidance of our investors rational investment philosophy, science investment decisions provides a theoretical basis.