Specification Tests for The Multinomial Logit Model Revisited: The Role of Alternative-Specific Constants

Jan Rouwendal

No 17-068/VIII, Tinbergen Institute Discussion Papers from Tinbergen Institute

Abstract: This paper shows that Hausman-McFadden and related tests for the multinomial logit model are not informative about failure of IIA, but only about specification error in the deterministic part of the utilities attached to choice alternatives. It is shown that the multinomial logit cannot be distinguished from any more general additive random utility model when a full set of alternative specific constants is present. Estimation of such more general models only addresses specification error in the deterministic parts of the utilities and is not informative about the true specification of the random terms of the utilities.

Keywords: multinomial logit; Hausman-McFadden specification test; nested logit; GEV models; mixed logit (search for similar items in EconPapers)
JEL-codes: D1 D4 (search for similar items in EconPapers)
Date: 2017-07-27, Revised 2018-01-29
New Economics Papers: this item is included in nep-ecm and nep-upt
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (2)

Downloads: (external link)
https://papers.tinbergen.nl/17068.pdf (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:tin:wpaper:20170068

Access Statistics for this paper

More papers in Tinbergen Institute Discussion Papers from Tinbergen Institute Contact information at EDIRC.
Bibliographic data for series maintained by Tinbergen Office +31 (0)10-4088900 ().