The paper summarized domestic and international research results in the field of Markowitz's mean-variance model. On the basis of these results, it proves the property of equivalence of between parameters' estimation of the minimum variance portfolio and that of the linear model, when the return sequences are strictly stationary. Also, it raises the translation invariance in the strictly stationary return sequences.Meanwhile, the estimated parameters with equation constrains are derived by using Lagrange multiplier law.Finally, an iterative algorithm of estimated parameters in the strictly stationary return sequences, which error terms are subject to AR(1) is put forward.
|