A0623
Title: Count data models with endogeneity and selection
Authors: Yves Croissant - Universite de La Reunion (France) [presenting]
Abstract: Endogeneity of some covariates in regression and endogenous selection are important problems in econometrics. Relevant methods of estimation have been proposed for linear models, namely instrumental variables, the general method of moments and the Heckman model for sample selection. More recently, these methods have been adapted to non-linear models. The contribution surveys the estimators proposed for count data: Muhally's GMM estimator for endogenous covariates, Terza's ML and NLS estimator for endogenous switching and Greene's ML and NLS estimator for endogenous selection, presents the implementation of these estimators in R and reproduce the original empirical examples contained in the corresponding articles (smoking habits and cigarette demand, credit card holding and major derogatory reports, physician advice and alcohol consumption, car ownership and mobility)
