Model identification and non-unique structure

Hendry, David F., Lu, Maozu and Mizon, Graham E. (2001) Model identification and non-unique structure (Economics Working Papers, 2002-W10) Oxford, GB. University of Oxford 16pp.

Record type: Monograph (Working Paper)

Abstract

Identification is an essential attribute of any model’s parameters, so we consider its three aspects of ‘uniqueness’, ‘correspondence to reality’ and ‘interpretability’. Observationally-equivalent overidentified models can co-exist, and are mutually encompassing in the population; correctly-identified models need not correspond to the underlying structure; and may be wrongly interpreted. That a given model is over-identified with all over-identifying restrictions valid (even asymptotically) is insufficient to demonstrate that it is a unique representation. Moreover, structure (as invariance under extended information) need not be identifiable. We consider the role of structural breaks to discriminate between such representations.

Text

DEISTGEM.pdf - Version of Record

Restricted to Repository staff only

More information

Published date: May 2001

Additional Information: Prepared for: Contributions to Econometrics, Time Series Analysis, and System Identification: A Festschrift in Honour of Manfred Deistler, Journal of Econometrics, edited by Benedikt M. P¨otscher and Ingmar R. Prucha.