Econometrica: Jan, 1964, Volume 32, Issue 1

Three-Stage Least-Squares and Full Maximum Likelihood Estimates

https://doi.org/0012-9682(196401/04)32:1/2<77:TLAFML>2.0.CO;2-2
p. 77-81

J. D. Sargan

This paper proves in the context of maximum likelihood estimation of linear stochastic models of the Cowles Commission type [2], that if the model is fully identified and stable and the error variance matrix unrestricted, three-stage least-squares estimates differ asymptotically from full maximum likelihood estimates by order 1/T, where T is the number of time periods. When the full maximum likelihood estimates are best asymptotic normal so are the three-stage least-squares estimates.