Font Size: a A A

Dynamic Spectrum Of Risk Measurement And Optimal Portfolio Analysis

Posted on:2012-07-21Degree:MasterType:Thesis
Country:ChinaCandidate:X Z ZhangFull Text:PDF
GTID:2219330338951093Subject:Statistics
Abstract/Summary:PDF Full Text Request
With the gradual opening of China's financial market and the change of economic cycling, China's financial market will become increasingly subject to international financial markets,and the volatility of China's financial markets is growing.However, due to China's financial system is not perfect, financial institutions and corporates'risk management and awareness of prevention is not mature enough, coupled with our financial regulatory system is not sound.Therefore, accurating measurement of financial risks and enhancing financial risk management is that China's financial system should be urgently resolved.The paper based on the nature of coherent risk measures, showed the defects of VaR in practical applications, focued on the research of the ES and spectral risk measures(SRM), and also introduced the methods of constructing risk spectrum function.The main study of paper is the spectral risk measure (SRM), SRM is a large class of coherent risk measure which is generated by the ES, ES is a special spectral risk measure.To calculate easily, during the process of studying dynamic SRM, we focued on the analysis a special case of SRM --- ES model, and builded dynamic SRM model based on GARCH model. Taking into account the characteristics of stock data's fat tail, the sequence distribution of the tail is fitted with the GED distribution, and through empirical analysis to compare the accuracy of VaR and ES models based on the normal distribution,student-t distribution and GED distribution.Eventually coming to the results that ES model based on GED distribution estimates risk much better.Finally, the paper uses SRM to establish the mean-M model to solve the problem of portfolio optimization in discrete state. Establishing the linear programming model of minimizing M which reduce the amount of the model, then illustrated by an example calculation of the portfolio optimization model in the application. The results show that each investor's attitude to the risk of portfolio is different, different investors have different risk tolerance, manifested in the model with different SRM values.They would make different investment decisions according to expected return and different attitude of risks. Overall, the greater risks the investors bear the more expected return. Risk measurement method using SRM provides a theoretical basis to diverse investors behavior of investment, which also has a practical significance.
Keywords/Search Tags:VaR, (CVaR)ES, Coherent risk measures, Spectral risk measures, investment portfolio
PDF Full Text Request
Related items