Influence Diagnostics for Bayesian Spatial Econometrics

This thesis develops new influence diagnostic methods for Bayesian spatial econometric models. Following the introductory chapter, Chapter 2 introduces key concepts that feature in this thesis and places its contributions within the context of broader literature. Chapter 3 introduces a new influence diagnostic method which assesses the impact of perturbing a dataset on its posterior distribution when considering a Bayesian spatial autoregressive model. Chapter 4 focuses on model selection and introduces a new influence diagnostic method to assess the sensitivity of a Bayes factor to small perturbations in the data. Chapter 5 provides a summary of the thesis and discusses some directions of future research.