Tail Risk Of Portfolio For Skew-Distributions

Risk measurement model has been an hot topic, in recent years,people have paid more and more attention to tail risks. For a long time, the more popular model of tail risk is value-at-risk and tail conditional expectations. However,these two models have their own advantages and disadvantages,so they are not suitable to the measure of tail risk. Tail conditional expectation solves the consistency problem of risk measures which value-at-risk can not solve,but it can not measure the tail variability when measuring the tail risks,while Tail conditional variance model can solve the above problems.In this thesis, we study the tail conditional expectation under the hypothesis of skew-normal distribution and skewed-t distribution, and we try to decompose the risk model of TCE and TCV. In this way,we can know the contribution situation of the single asset’s risk for the whole portfolio risk in the portfolio. Then we consider the parametric sensitivity analysis of the TCE and TCV model,explore each parameter’s sensitivity in the models of tail risk model respectively, from many point of view to study the tail risk. Investors can adjust the model parameters related to the portfolio, so as to ensure that the risk models values of loss risk functions should be below a certain preset controllable level.