Model Validation in Spatial Econometrics: A Review and Evaluation of Alternative Approaches

Abstract

Formal approaches to assess the validity of spatial theoretical constructs are reviewed and evaluated with respect to their application and potential in spatial econometrics. A distinction is made between specification tests and model selection procedures. For the first, particular emphasis is on tests for residual spatial autocorrelation, tests on common factors, and tests on non-nested hypotheses. For the second, attention is focused on information-theoretic criteria, Bayesian approaches, and heuristic procedures. Also, some thoughts are presented on model validation as a multiobjective decision problem and its relevance in planning and policy analysis.