Application Research Of Financial Risk Measurement Method

In the 1970 s since the collapse of the Bretton woods system, floating exchange rate system of the international monetary system gradually established. The fluctuation of exchange rate and price factor interest rate is frequent, and the width of fluctuation in larger and larger. The influx of hot money in the global range exacerbate this kind of phenomenon, leading to a global economic and financial crisis, for individual investors, financial institutions and even the entire financial industry poses a serious threat. As a result, most financial institutions and organizations are very attaches great importance to the development and application of risk measurement model, hope to be able to accurate measurement and evaluation of the financial risk, in order to take reasonable avoidance measures, reduce the risk of financial losses.This paper detailed introduces the measurement method of the traditional financial risk nominal value measurement method, measurement sensitivity, volatility measurement method and VaR(VaR method is an extension of the method has many, here is differ a list).These methods exist some disadvantages, such as nominal value measurement method is simple, but obviously overestimated the financial risk; Measurement method to calculate the sensitivity coefficient can only be compared with the same portfolio of data; Volatility measurement method can not give the size of the loss value, and it is the same as the traditional VaR method has the disadvantages of the sequence is to assume that yields are to obey normal distribution, however, the financial product yield sequence can often be rush fat-tailed features, do not obey normal distribution.In this paper, on the basis of previous studies, the study of traditional measurement method of the financial market risk and financial risk measurement methods, comparing the advantages and disadvantages of the calculation method of VaR respectively yield mapping method and the standard history simulation method is introduced. After the main of VaR calculation method based on copulas connect-Garch model theory is introduced and the research in detail, copulas connect- Garch model is more flexible in the serial hypothesis, not only for its model and data fitting is better, more can describe the characteristics of financial return data Gao Fenghou tail.Finally, by the currency of the Internet financial products, the risk of Shanghai Pudong Development Bank and Zoomlion measure the empirical analysis, respectively introduces the traditional financial risk measurement methods and the advantages and disadvantages of copulas connect- Garch model to measure Va R method.In this paper, the conclusion has the following points. First the mean measurement method is applicable to the nominal value of extreme risk aversion of investors, and the calculation method is simple; Second, sensitivity measurement method to measure in a certain extent between the investment products, but its application has limitations, only in the portfolios of contrast, cannot be compared with the rest of the investment products; Three, volatility measurement method can measure the volatility of yield, but it cannot measure changes direction, nor given the loss of specific values; Four, traditional VaR method is not only to break the limitations of the above several methods to measure the specific loss value, it’s comparable, and given the probability of loss value, intuitive and easy to understand as a result, the calculation method is relatively simple, it is also one of the reasons for its mainstream risk measure. However, in the process of calculation, the traditional VaR method need to assume that the original distribution is normal distribution or completely conform to the historical distribution, this limitation brought questions to the accuracy of the VaR method; Five, Copulas- Garch model can very good fitting the sequence of yield of the portfolio, the use of copulas connect model need no assumption of the distribution of data, so in the same portfolio, allowing the existence of different distribution, makes the VaR application more widely, multivariate copulas connect function itself, the application of monte carlo simulation, you can get some data missing or unfavorable situation.