Font Size: a A A

Portfolio Investment In The Mean - Variance Model Parameter Estimation To Improve And Empirical Research

Posted on:2005-08-29Degree:MasterType:Thesis
Country:ChinaCandidate:J F LiFull Text:PDF
GTID:2206360125463997Subject:Quantitative Economics
Abstract/Summary:PDF Full Text Request
Foreign study about making modern portfolio investment decision has more than half century, as the basis of modern portfolio theory, mean-variance portfolio theory still has its usefulness in the theory study and application; it still is worth of studying.Mean-variance portfolio theory is one problem of quadratic programming, the traditional study about it pays more attention on the unbiased estimation about the two input parameters and loosening the constraints, and tries to break through the algorithm about this optimal problem. Mean-variance portfolio theory is not widely used in practice, because the model is sensitive to the change of two parameters, which will lead to the lack of stability of the portfolio structure and get the inferior performance out of sample period. All of those do not mean the theory is not perfect; the key is that the traditional estimation methods have estimation risk, which leads investors to choose the suboptimal portfolio.To solve such problem, several methods can decrease estimation risk are given in the paper. We only study the application of the Stein-rule in the field of portfolio theory and empirical study in this paper.(1) We shrink the mean of the mean-variance portfolio theroy by Stein-rule and compare the use of James-Stein rule to the use of Bayes-Stein rule for improving the performance of portfolio. At the same time, we study the portfolio with constraints of investment weight. The result shows that better performance can be gained by improving mean parameter estimation.(2) After shrinking the covariance matrix of the mean-variance portfolio theory by Stein-rule and giving one simple method of calculating the optimal shrinkage intensity of covariance matrix, we compare the standard deviation, variable coefficient and the forecasting capability of different minimum variance portfolio defined by different covariance matrix. The result shows that covariance matrix improved by Stein-rule can gain better performance.(3) After shrinking the investment weight of the mean-variance portfolio theoryby Stein-rule, we make progress on the problem of how to judge the weight of the combination of the active portfolio management and passive portfolio management which is noticed by other scholars.
Keywords/Search Tags:portfolio, Stein-rule mean, covariance matrix, investment weight.
PDF Full Text Request
Related items