SENSITIVITY OF THE RESTRICTED LEAST SQUARES ESTIMATOR UNDER MISSPECIFIED RESTRICTIONS AND MULTICOLLINEARITY

Ingrid Kreuzmair (Germany) and Helge Toutenburg (Germany)

Received September 6, 2007

Abstract

The availability of prior information in the form of exact linear restrictions is utilized in the estimation of the parameters of a linear regression model. The ordinary least squares estimator (OLSE) and the restricted least squares estimator (RLSE) are obtained and compared under the criterion of the MSE matrix when the prior information is misspecified and when multicollinearity is present in the data. The dominance condition is numerically analysed and guidelines are proposed.

Keywords and phrases: linear restrictions, misspecification, restricted regression estimator, mean squared error matrix superiority, multicollinearity, simulation.