Introduction to Spatial Econometrics

Although interest in spatial regression models has surged in recent years, a comprehensive, up-to-date text on these approaches does not exist. Filling this void, Introduction to Spatial Econometrics presents a variety of regression methods used to analyze spatial data samples that violate the traditional assumption of independence between observations. It explores a wide range of alternative topics, including maximum likelihood and Bayesian estimation, various types of spatial regression specifications, and applied modeling situations involving different circumstances. Leaders in this field, the authors clarify the often-mystifying phenomenon of simultaneous spatial dependence. By presenting new methods, they help with the interpretation of spatial regression models, especially ones that include spatial lags of the dependent variable. The authors also examine the relationship between spatiotemporal processes and long-run equilibrium states that are characterized by simultaneous spatial dependence. MATLAB® toolboxes useful for spatial econometric estimation are available on the authors’ websites. This work covers spatial econometric modeling as well as numerous applied illustrations of the methods. It encompasses many recent advances in spatial econometric models—including some previously unpublished results.

Introduction Spatial dependence The spatial autoregressive process An illustration of spatial spillovers The role of spatial econometric models The plan of the text Motivating and Interpreting Spatial Econometric Models A time-dependence motivation An omitted variables motivation A spatial heterogeneity motivation An externalities-based motivation A model uncertainty motivation Spatial autoregressive regression models Interpreting parameter estimates Maximum Likelihood Estimation Model estimation Estimates of dispersion for the parameters Omitted variables with spatial dependence An applied example Log-Determinants and Spatial Weights Determinants and transformations Basic determinant computation Determinants of spatial systems Monte Carlo approximation of the log-determinant Chebyshev approximation Extrapolation Determinant bounds Inverses and other functions Expressions for interpretation of spatial models Closed-form solutions for single parameter spatial models Forming spatial weights Bayesian Spatial Econometric Models Bayesian methodology Conventional Bayesian treatment of the SAR model MCMC estimation of Bayesian spatial models The MCMC algorithm An applied illustration Uses for Bayesian spatial models Model Comparison Comparison of spatial and non-spatial models An applied example of model comparison Bayesian model comparison Chapter appendix Spatiotemporal and Spatial Models Spatiotemporal partial adjustment model Relation between spatiotemporal and SAR models Relation between spatiotemporal and SEM models Covariance matrices Spatial econometric and statistical models Patterns of temporal and spatial dependence Spatial Econometric Interaction Models Interregional flows in a spatial regression context Maximum likelihood and Bayesian estimation Application of the spatial econometric interaction model Extending the spatial econometric interaction model Matrix Exponential Spatial Models The MESS model Spatial error models using MESS A Bayesian version of the model Extensions of the model Fractional differencing Limited Dependent Variable Spatial Models Bayesian latent variable treatment The ordered spatial probit model Spatial Tobit models The multinomial spatial probit model An applied illustration of spatial MNP Spatially structured effects probit models References A summary appears at the end of each chapter.

James LeSage, Robert Kelley Pace