Optimization Model Of Financial Market Risk Measurement

In last a decade, financial service has entered into an era with complexity and full of mathematical challenge. A series of financial events and crises have threatened severely the existence and development of financial institution. Thus risk measurement technique has grown into a great concern of financial service and relative academic.The paper has four chapters. Chapter 1 gives an overview about the meaning, cause, type of financial risk as well as the present situation on its measurement.Chapter 2 provides evidences that the CAPM model does not work in Shanghai stock market. In order to make up the CAPM model, a test of the conditional CAPM using the GARCH-M framework is conducted and existing research is extended by investigating the effect of intervals by conducting tests over different sampling frequencies. Not only the GARCH-M model but also the conditional CAPM are supported for daily and weekly return intervals while greater support for the model is found in the daily return intervals.Chapter 3 introduces the basic models and algorithms of prevail used risk measurement method-Value at Risk (VaR). Also Back-tests of the models are checked and comparisons between them are investigated. Then Chapter 3 provides evidences from China's stock market that Estimating Functions model and GARCH-M model are fitted and verified respectively. Further, a modified model is given by alternatively using these two models to enhance the previous models of VaR. It is pointed out that the modified one is practical and useful in application of risk managment.Assuming that the rates of return obey normal distribution, Chapter 4 combines the advantages between the mean-VaR model and the chance-constrained programming model and presents a chance-constrained mean-VaR portfolio problem with short selling, which is determined by expected rate of return and confidence level. Its mathematical model is established, and the properties of existence anduniqueness of the optimal solution are discussed. Furthermore, the explicit representation of the optimal solution is given. By Matlab language, the program of obtaining the optimal solution is devised. Finally, an illustrative example is provided and two conclusions are proved.