Semiparametric estimation of count regression models

摘要: This paper develops a semiparametric estimation approach for mixed count regression models based on series expansion the unknown density of unobserved heterogeneity. We use generalized Laguerre around gamma baseline to model heterogeneity in Poisson mixture model. establish consistency estimator and present computational strategy implement proposed techniques standard as well truncated, censored, zero-inflated models. Monte Carlo evidence shows that finite sample behavior is quite good. The applies method individual shopping behavior.