Recent developments in spatial stochastic frontier models

This chapter discusses the empirical literature that uses a spatial stochastic frontier (SSF) analysis framework in production economics. We first outline standard spatial extensions of the classical stochastic frontier models aiming to capture the spillover effects of neighbouring production units. We then turn our attention to studies mainly focused on spatially correlated noise and inefficiency terms. Following this, semiparametric and nonparametric models, and modern approaches incorporating space-varying coefficients, unobserved common factors, stochastic spatial weight matrices, endogeneity and latent class structures, are discussed. We then provide a section with guidelines for researchers regarding model selection followed with a section devoted to issues relating to returns to scale and productivity growth are then discussed. A brief guide for practitioners to available software is provided in an appendix.