Semiparametric estimation of on-stie count data models

This article proposes semiparametric estimation of on-site count data models based on a series expansion approach of Gurmu, Rilstone and Stern (1999, Journal of Econometrics 88, 123-150), which is flexible and adaptable for a form of overdispersion as long as the exponential mean parameterization is given. We also provide the empirical illustration of demand for a recreation site. The result suggests that the existing parametric approaches will cause wrong statistical inference for the on-site count data.