A Spatial Sample Selection Model*

Abstract

This paper presents a sample selection model with spatial correlation in the selection and outcome variables and studies the maximum likelihood method of estimation. Consistency and asymptotic normality of the maximum likelihood estimator are established by the spatial near‐epoch dependent properties of the variables. Monte Carlo simulations show its good finite‐sample performance. This model is used to examine the impact of climate change on cereal yields in Southeast Asia and projects that climate change may cause a reduction in cereal yields by () in the minimum‐change (maximum‐change) scenario.