Estimation of spatial sample selection models: A partial maximum likelihood approach

Abstract

We study estimation of sample selection models with the spatially lagged latent dependent variable or spatial errors in both the selection and outcome equations under cross-sectional dependence. Since there is no estimation framework for the spatial-lag model and the existing estimators for the spatial-error model are computationally demanding or have poor small sample properties, we suggest to estimate these models by the partial maximum likelihood estimator. We show that the estimator is consistent and asymptotically normally distributed. To facilitate easy and precise estimation of the variance matrix, we propose the parametric bootstrap method. Simulations demonstrate the advantages of the estimators.