The Down-sided Risk Measure Models In Stock Market And The Demonstration Research About The Market Risk

Searching the method of risk measure in the stock market is a very important work in finance. More and more investor have payed more and more attention in this field. Since the appearance of Markowitz(1952)'s assets portfolio theory, the research of risk measure not only stay on the subjective judgement, but also discuss and analyse it from the ration,the psychology and direction of risk ect. There are many useful conclusions in this interesting field.This article mainly researches the down-side risk models of the stock market and their application in China's stock market. Researching risk measure models is one of the significant problems in finance field and is a very important process in actual investment. So there are many people make great efforts to settle this problem. But many current models just think of the rate of return and not take care of the endure ability of the investor. Then we firstly sum up most risk measure models and analyse their merit and shortcoming. Secondly, in the third and forth chaper we take into account the preference of investor to bring forward a LVaR model on the base of MGARCH model and the LCVaR model. Fianally, in the fifth and sixth chaper we distinguish the"loss"and"profit"of down-side risk of the inbestor and build the all-sided risk measure and the compounding pre-quadrature model.Now, we give the main content of every chapter of this article:The first chapter is the exordium of the literature. It expatiate the definition character and the class of risk in stock market. And introduce the framework and innovation of this literature. The first section analysis the shortcomings of some risk measure. The second section mainly expatiate the definition and character of risk in stock market which is the basement of our researching object—risk in stock market. The third section introduce the framework and innovation of this literatureIn second chapter, we differentiate the risk measure models to two theorys: one is partial to pure theory and the others are partial to practicality. In the same time, the latter can be distinguished to two types: models based on the indetermination and down-side risk measure models. So the main objective of this chapter is to introduce these models and appraise them. In the first section, we introduce the risk measure models partial to pure theory. The three types of it are models based on the expectation utility function, the standard risk measure models and the Stochastic Dominance theory. The main content of the second section is the models partial to practicality. Indetermination risk theory includes the variance,βvalue, H index, information entropy etc and the down-side risk measure include the down-side pre-quadrature, half-absolute deviation, VaR, CVaR, common risk measure models and risk manage based on the behavior finance. Finally we evaluate the merit and the disadvantage of all these models. The third section mainly introduces some researching conclusion of China's scholar.The main content of the third chapter is build the LVaR model based on the MGARCH model. This model can describe the fat-tail phenomenon better when compares with the normal GARCH model. The first section introduce the definition and form of MGARCH model, tail character and the calculate method of the parameter. In the second section, we introduce the definition and form of LVaR model and build a LVaR model based on the MGARCH model. We use the EM algorithm to calculate this model and discuss the proof-test and the choice method of this model. The third section is demonstration part. After comparing the result, we can think the model based on MGARCH model consider the fat-tail phenomenon much little more than others. So it will little better than the models based on the GARCH model. The result tell us that the LVaR based on MGARCH model can portray the"high apex and deep trail"phenomena better than GARCH models. Because the former think of the trail character of sotck market much more. The main content of the fourth section improve on the Rockafellar's CVaR model -that's the LCVaR model-from risk preference of every investor. Then we ues this model to test China's stock market risk. The first section describes the merit of coherent risk measure which called CVaR and compares its character with VaR model. We can say, the former is better than the latter. In the same time, we discuss the calculate method of LCVaR model. In the second section we provide a portfolio optimization model based on the LCVaR model. Last section is demonstration part and we apply the LCVaR model into shanghai stock market and give out some conclusions. When we fix on the goal profit of the investor, we can find out that their tolerance capability is different and the goal is higher and the risk is bigger. From the efficacious border of the MV and Mean-LCVaR, although they are very similar, their invest proportion is different. That's because LCVaR model take care of the psychology of investors and it's a better optimization model.In the fifth chapter, on the base of down-side risk measure model, we introduce the risk adjusted factor M to build the all-sided risk measure model. The first section describes the definition,form and calculate of positive deviate and the negative deviate. Then we discuss four kinds of models: model based on the intensity and probability of the negative deviate, model based on the size of pressure, model based on fluctuate frequency and the model which integrate three models. In the second section, we build the compounding pre-quadrature model and then build the all-sided risk measure model which include the the positive deviate adjusted risk factor. In last section, we compare the compounding pre-quadrature model and the all-sided risk measure model with the demonstration in China's stock market. In the demonstration part, we find out that the risk from the full-side risk measure model is bigger than the result from compounding pre-quadrature model. That's because the former are composed with more risk factors than the latter.The sixth chapter is continuity of fifth chapter. We need to resolve portfolio optimization problem. After we discussed the relation of portfolio risk and one stock risk, we introduce the portfolio optimization model based on the all-sided risk measure model and its availability boundary. They are the supposition for portfolio optimization in the first section. In the second section, we first discuss the relation of portfolio risk and one stock risk. Then, under the restrict of expectation assets return rate, we provide the form of all kinds of portfolio optimizations which to minimize the R1 , R 2, R3 , R 4and the all-sided risk measure model. In the third section we discuss the method of calculating the portfolio optimization models and their availability boundary. In the demonstration part, we find out that the efficacious border of the two models is also similar and the results are consistent with the conclution in the forth section. But the invest proportion is very dissimilar with the result from MV model. That's because the former three take into account of the preference of different investors whereas the MV model mix the concept of"loss"and"profit"which not accord with the quality of the risk of the stock market.