An equation by any other name is still the same: on spatial econometrics and spatial statistics

Abstract

Statistics is a branch of mathematics concerned with the collection, quantification, analysis, interpretation, and presentation of real-world data, and the use of probability theory to estimate population parameters with these data. Spatial statistics is a subset of statistics that is concerned with handling the special problems associated with geographically distributed data, which include spatial point patterns, regional and lattice measurement aggregations, irregularly spaced site-specific measurements on a surface, and image analysis. Meanwhile, econometrics is concerned with the application of statistical methods to the study of economic data and problems. When coining the term spatial econometrics in 1979, Paelinck and Klaassen characterized it as a subset of econometrics that is concerned with the role of spatial dependence in regional economic model response and explanatory variables, asymmetries in spatial relationships, the specification of geographic structure governing spatial interactions, and the explicit modeling of space. We outline and discuss principal similarities (e.g., testing for the presence of spatial autocorrelation) and differences (e.g., map generalization) between spatial statistics and spatial econometrics. In doing so, our goal is to help clarify past, present, and future relationships between these two subfields.