A Note on Two-Stage Least Squares, Three-Stage Least Squares and Maximum Likelihood Estimation in an Expectations Model

An attempt is made to exploit the similarities between the limited information model and the expectations model. Given that the 2 models differ in only one crucial aspect, it is tempting to use the methods of estimation associated with the limited information model. However, the danger of extending...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:International economic review (Philadelphia) 1985-06, Vol.26 (2), p.507-510
1. Verfasser: Turkington, Darrell A.
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:An attempt is made to exploit the similarities between the limited information model and the expectations model. Given that the 2 models differ in only one crucial aspect, it is tempting to use the methods of estimation associated with the limited information model. However, the danger of extending results obtained in the limited information model to models that are extremely similar is demonstrated. In the expectations model, assuming right hand exogenous variables are not used in forming expectations, 2-stage least squares (2SLS) is not as efficient asymptotically as maximum likelihood estimation (MLE), and 3-stage least squares (3SLS) differs even asymptotically from 2SLS. In a specific case, 2SLS is less efficient than the 2-step estimator. However, if asymptotically efficient estimators of the parameters of the model are desired, MLE or 3SLS should be applied.
ISSN:0020-6598
1468-2354
DOI:10.2307/2526598