Spectral Measure Of Risk Analysis Based On G-H Distribution And Copula-SMR Method

In financial markets, investors and financial institutions are facing moreserious financial risks by market fluctuations. How to accurately measure themarket risks, to help investors and managers of financial institutions at risk, isthe main research work of economists. This paper choose spectrum riskevaluation model, and introduce G-H distribution, to be the main riskmeasurement method in financial markets.Risk spectrum function is used to reflect the degree of risk aversion toinvestors. There are three main forms: exponential risk spectrum function,power risk spectrum function and hyperbolic risk spectrum function. In thispaper we use exponential and hyperbolic risk spectrum function, choose G-Hdistribution, combine with the spectrum risk measurement method, to getinvestment portfolio measure model, and use the trapezoidal method to getdiscrete form respectively.In this paper, we select the KangMei pharmaceutical yield sequence assample data to do the single stock risk measurement analysis. We use quantileestimation method and moment estimation methods to do G-H distributionparameter estimation respectively. Compared with the result of the riskspectrum measurement model based on normal distribution, we get conclusion:elected the G-H distribution and moment estimation method, the fitting effect isthe best. In certain confidence level of sample data, we choose Kupiec failureinspection to do risk spectrum measurement analysis, while compare with thedata out of sample, we get the number of forecast failure, and get conclusion.In this paper we choose Shanghai home, etc. four yields to do investmentportfolio risk measurement analysis. We use exponential and hyperbolic riskspectrum function to get investment portfolio risk measure model, and getconclusion: with the increase of the expected yield or absolute risk aversionfactor, risk spectrum measurement values increase accordingly, and weight will transfer to the stock with high average yield.Also we use Copula function to measure correlation between differentsecurities. Select Wanke A and conch cement, choose proper Copula functionto establish investment portfolio model. The empirical results show that,because of copula function considering the correlation of the end of thesecurities, it can get the results that more close to the actual market condition.