Scale-dependence of the Negative Binomial Pseudo-Maximum Likelihood Estimator

Clément Bosquet (1) , Hervé Boulhol (2, 3)

1 GREQAM - Groupement de Recherche en Économie Quantitative d'Aix-Marseille
2 OCDE - Organisation de Coopération et de Développement Economiques = Organisation for Economic Co-operation and Development
3 CES - Centre d'économie de la Sorbonne

Following Santos Silva and Tenreyro (2006), various studies have used the Poisson Pseudo-Maximum Likelihood to estimate gravity specifications of trade flows and non-count data models more generally. Some papers also report results based on the Negative Binomial estimator, which is more general and encompasses the Poisson assumption as a special case. This note shows that the Negative Binomial estimator is inappropriate when applied to a continuous dependent variable which unit choice is arbitrary, because estimates artificially depend on that choice.

Mots clés: pseudo-maximum likelihood methods; negative binomial estimator; Poisson regression; gamma PML

Domaines: Economies et finances

Clément Bosquet, Hervé Boulhol. Scale-dependence of the Negative Binomial Pseudo-Maximum Likelihood Estimator. 2010. ⟨halshs-00535594⟩