This volume focuses on econometric models that confront estimation and inference issues occurring when sample data exhibit spatial or spatiotemporal dependence. This can arise when decisions or transactions of economic agents are related to the behaviour of nearby agents. Dependence of one observation on neighbouring observations violates the typical assumption of independence made in regression analysis. Contributions to this volume by leading experts in the field of spatial econometrics provide details regarding estimation and inference based on a variety of econometric methods including, maximum likelihood, Bayesian and hierarchical Bayes, instrumental variables, generalized method of moments, maximum entropy, non-parametric and spatiotemporal. An overview of spatial econometric models and methods is provided that places contributions to this volume in the context of existing literature. New methods for estimation and inference are introduced in this volume and Monte Carlo comparisons of existing methods are described. In addition to topics involving estimation and inference, approaches to model comparison and selection are set forth along with new tests for spatial dependence and functional form. These methods are applied to a variety of economic problems including: hedonic real estate pricing, agricultural harvests and disaster payments, voting behaviour, identification of edge cities, and regional labour markets. The volume is supported by a web site containing data sets and software to implement many of the methods described by contributors to this volume
