Working Papers Home

2015 Working Papers
2014 Working Papers
2013 Working Papers
2012 Working Papers
2011 Working Papers
2010 Working Papers
2009 Working Papers
2008 Working Papers
2007 Working Papers
2006 Working Papers
2005 Working Papers
2004 Working Papers
2003 Working Papers
2002 Working Papers
2001 Working Papers
2000 Working Papers

Search All Papers

JEL Classification

Past Working Papers (Prior to 2000)

Office of Research
Home Page

Information on
Submitting a Paper

"On Some Heteroskedasticity-Robust Estimators of Variance-Covariance Matrix of the Least Squares Estimators"

Anil K. Bera, Totok Suprayitno, and Gamini Premaratne


First Author :

Anil K. Bera
University of Illinois at Urbana-Champaign
1206 S. Sixth Street, M/C 706
Champaign, IL 61820

Second Author :

Totok Suprayitno
Research Scientist
Ministry of Education, Indonesia

Third Author :

Gamini Premaratne
University of Illinois at Urbana-Champaign
1206 S. Sixth Street, M/C 706
Champaign, IL 61820

Create Profile

Abstract :
Chesher and Jewitt (1987) demonstrated that the Eicker (1963) and White (1980) consistent estimator of the variance-covariance matrix in heteroskedastic models could be severely biased if the design matrix is highly unbalanced. In this paper we, therefore, reconsider Rao’s (1970) minimum norm quadratic unbiased estimator (MINQUE). We derive the analytical expressions for the mean squared errors (MSE) of the Eicker-White, one of MacKinnon and White’s (1985) and MINQUE estimators, and perform a numerical comparison. Our analysis shows that although MINQUE is unbiased by construction, it has very large variance particularly for the highly unbalanced design matrices. Since the variance is the dominant factor in our MSE computation, MINQUE is not the preferred estimator in terms of MSE comparison. We also studied the finite sample behavior of the confidence interval of regression coefficients in terms of coverage probabilities based on different variance-covariance matrix estimators. Our results indicate that although MINQUE generally has the largest MSE, it performs relatively well in terms of coverage probabilities.
Manuscript Received : 2000
Manuscript Published : 2000
This abstract has been viewed 3247 times.