Spatial Econometrics: Statistical foundations and applications to regional convergence

An increasing number of books have been devoted to spatial econometrics over the last few years.The trend is actually very similar for articles based on this technique; more and more are published in renowned regional science and geography journals each year.So one may wonder what makes Giuseppe Arbia's book so special compared to other contributions in the field of spatial econometrics.I would say the competitive advantage of his book is twofold.First, the book starts with a very explicit chapter devoted to introducing the reader to the idea that the distribution of per capita GDP is not random across regions, and thus Ordinary Least Squares estimations are not necessarily appropriate when applied to spatial data.Even a reader who never heard a word about spatial econometrics will find it clear and hopefully be interested in the rest of the book.Chapters 2 to 4 are the second strong point of this book.They review in a rather rigorous framework all the statistical concepts and tools at the basis of spatial econometric modelling.In this sense, the subtitle of Arbia's book, "statistical foundations and applications to regional convergence", is more descriptive of its content than the title itself.The book has a clear structure with a first chapter highlighting how the geographical component of spatial datasets needs special econometric treatments.He draws his conclusions from an OLS estimation of the absolute betaconvergence process measured across both EU and Italian regions.Arbia does not return to his empirical example before chapter 5 where he formally performs a spatial econometric estimation of the convergence process on his datasets.In the meantime, chapter 2 provides an introduction to random fields and the restrictions that need to be considered to estimate a statistical linear regression model based on spatial data.Chapter 3 starts by introducing the basic notation and concept of likelihood, its related testing procedure, and then the discussion moves to the