Why Do We Reject the Mean-Variance Model?
提出一个两阶段检验程序，用于静态均值-方差投资组合模型，并应用于德国私人净财富的季度数据，发现模型被强烈拒绝，原因在于数据否定了收益与资产持有之间的任何正半定关系。
A two-stage test procedure for the static mean-variance portfolio model is presented which focuses on increasingly restrictive versions of the general model. It also delivers clues to the probable cause of rejection. The procedure is applied to quarterly data for German private net wealth. The mean-variance restrictions are strongly rejected because the data reject any positive semidefinite link between returns and asset holdings. This suggests that the failure of the model is not primarily caused by neglect of the time variation of conditional covariance. Copyright 1995 by The editors of the Scandinavian Journal of Economics.