ivprobit-1.0: Instrumental Variables Probit Model

title: 'ivporbit:An R package to estimate the instrumental variables probit model' tags: - limited-dependent variable models - Amemiya's Generalized Least Squares estimators - R - ivprobit authors: - name: Taha Zaghdoudi orcid: 0000-0002-4488-5774 affiliation: 1 affiliations: - name: Faculty of Law, Economics and Management of Jendouba, Tunisia index: 1 date: 13 December 2017 bibliography: paper.bib

Summary

ivprobit [@zaghdoudi] is an open source R package that fit the instrumental variables probit model. The package comprises function that deal with endogeneity misspecification especially when correlation between regressors and error term is determinited and which produce inconsistent results. To avoid this problem, authors applied the instrumental variable method in which they use the correlated variables as instruments. Nevertheless, a two-stage method are used by some authors [@blundell2004endogeneity] to fit the probit model but it produce non efficient result.

However, [@newey1987efficient] expose an efficient way to estimate limited-dependent variable model by using the Amemiya's Generalized Least Squares estimators (AGLS) [@Amemiya1978]. Likewise, in the ivprobit package we use the AGLS method and we include in the two-stage model a continuous endogenous regressor. This method is used when the maximum likelihood method (MLE) fail to estimate the model.

ivprobit is GPL-3 licensed and can be retrieved from GitHub Repository

ivprobit CRAN