Research On Optimal Multi-stage Portfolio Selection With Background Risk

Portfolio selection has been a popular topic in financial flied.It is concerned with distribute one's initial wealth to different assets in the capital market in order to achieve the optimal portfolio decision.In 1952,Markowitz published "portfolio selection",In this paper,he used quantitative analysis on the research of portfolio selection for the first time,and he established the mean-variance model.Later,more and more scholars begin to study portfolio theory deeply.The classical mean-variance model is extended to various forms of portfolio models.However,most models are constructed based on investors only face the financial risks caused by the fluctuation of asset prices.In the actual investment process,the investors not only face the financial risks,but also non-financial risks that cannot be dispersed through asset allocation,such as those caused by labour income,real estate investment and health.This kind of risk is called background risk.Some achievements have been made in portfolio research considering background risk,but the current research on background risk is based on single-stage investment decision.In this thesis,we considering background risk in multi-stage portfolio selection problem,and under the framework of mean-variance model,we established a multi-stage mean-variance model based on background risk.Then,we solved the model by using Lagrange duality method and dynamic programming method.we obtained the optimal strategy of multi-stage mean-variance model considering background risk and the portfolio efficient frontier.Finally,we compared a multi-stage mean-variance model with background risk with a multi-stage mean-variance model without background risk.The study suggests that the background risk does influence the investor's investment decisions in the multi-stage investment process,and at a given expectation of the investor terminate wealth,consider The risk of the optimal portfolio of background risk is higher than that of the optimal portfolio without considering the background risk.