Bayesian Estimation of Seemingly Unrelated Regression with Collinear Categorical Explanatory Variables

Abstract

The efficiency of Seemingly Unrelated Regression (SUR) not only depends on contemporaneous correlation between errors term but also on the degree of collinearity among explanatory variables in the equations of the model. The problem of collinearity consequentially leads to biased estimation, rank deficient, large standard deviations and misleading interpretation of the estimates among others in analysis. This study examines the robustness of the Bayesian estimator to varying degree of correlation among categorical explanatory variables of seemingly unrelated regression model. The result revealed an asymptotic property of the Bayesian method and the best estimates were obtained when sample size, N is large irrespective of the degree of correlation among the regressors.

Keywords: Bayesian estimator, Seemingly Unrelated Regression, collinearity, posterior standard deviations

Article Review Status: Published

Pages: 1-8 (Download PDF)

Creative Commons Licence
This work by European American Journals is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License