Home  >>  Archives  >>  Volume 14 Number 4  >>  st0359

The Stata Journal
Volume 14 Number 4: pp. 778-797

Subscribe to the Stata Journal

dhreg, xtdhreg, and bootdhreg: Commands to implement double-hurdle regression

Christoph Engel
Max Planck Institute for Research
on Collective Goods
Bonn, Germany
Peter G. Moffatt
School of Economics
University of East Anglia
Norwich, UK
Abstract.  The dhreg command implements maximum likelihood estimation of the double-hurdle model for continuously distributed outcomes. The command includes the option to fit a p-tobit model, that is, a model that estimates only an intercept for the hurdle equation. The bootdhreg command (the bootstrap version of dhreg) may be convenient if the data-generating process is more complicated or if heteroskedasticity is suspected. The xtdhreg command is a random-effects version of dhreg applicable to panel data. However, this estimator differs from standard random-effects estimators in the sense that the outcome of the first hurdle applies to the complete set of observations for a given subject instead of applying at the level of individual observations. Command options include estimation of a correlation parameter capturing dependence between the two hurdles.
Terms of use     View this article (PDF)

View all articles by these authors: Christoph Engel, Peter G. Moffatt

View all articles with these keywords: dhreg, xtdhreg, bootdhreg, hurdle model, double-hurdle model, random-effects double-hurdle model, tobit, p-tobit, inverse Mills ratio, bootstrapping

Download citation: BibTeX  RIS

Download citation and abstract: BibTeX  RIS