RESEARCH PAPER NO. 794
SAMPLE SIZE REQUIREMENTS FOR ESTIMATION IN SUR MODELS
by
WILLIAM E. GRIFFITHS, CHRISTOPHER L. SKEELS & DUANGKAMON CHOTIKAPANICH
JUNE 2001
Department of Economics. University of Melbourne. Melbourne Victoria 3010 Australia
ABSTRACT
This paper explores sample size requirements for the estimation of SUR models by (two-stage) feasible generalized least squares, maximum likelihood and Bayesian methods. It is found that the sample size requirements presented in standard treatments of SUR models are incomplete and potentially misleading. It is also demonstrated that likelihood-based methods potentially require larger sample sizes than does the two-stage estimator considered in this paper.
This mail form will be sent automatically to Department of Economics Administrative Staff and your request will be processed. Please fill in your Details below and Click On "Send Request" button when you are finished. Papers will be sent by post. |