Hurdle models for overdispersed count data

Hurdle models are frequently used to model count data. In this proof of principle study, I propose two extensions of hurdle models that make popular specifications more flexible. Both extensions nest the models which have been used so far. An example from health economics illustrates the relevance of these extensions. To foster the use of these models, I provide three user-written Stata commands. In a current project I analyze the performance of nonparametric count data models in small samples.