Spatial Econometrics

Spatial econometrics is a subarea of the broader field of econometrics. It can be defined as the branch of economics that applies statistical and mathematical rigor to the study of economic theories and relationships. By bringing econometric theory and methods to the analysis of spatial data, spatial econometrics offers an appropriate analytical framework and approach in situations where traditional statistical tools and methodologies often are inappropriate.

A spatial econometric approach to regional variations in income, for example, could be to ask why household income is higher in some regions of the [Page 410]country than in others. The spatial econometrician might begin by examining, say, median income of households from a national census, for all subregions of a country. The income variable might then be modeled by taking account of other variables for each sub-region, such as the average education level of adults, the unemployment rate for persons of working age, and other variables believed to be reasonable predictors of income. Such data are termed cross-sectional, because they refer to attributes of several units of geography for a single point in time. Cross-sectional data for multiple time points (panel data) are an extension of such analyses into the space-time domain. In addition to attribute values, however, no matter how rich and detailed they may be, a spatial data set is of interest to an applied spatial econometrician only when the location of each subregion is recorded, becomes part of the data set, and plays a role in the analysis of the data. Spatial econometric analysis, therefore, seeks to understand relationships among attribute values but only by taking full account of the location of the geographic units from which the attribute data are taken.

The field of spatial econometrics also can be viewed as a subarea of the broader, more general field of spatial analysis. Two generalizations can be made in this regard. First, spatial econometrics usually involves the analysis of spatial area objects or polygons. When mapped, such areas constitute what is referred to as a spatial lattice. A spatial lattice might consist of all census tracts in a county, all counties or other major subdivisions of a country, or all countries in a major region of the world. Such a lattice is called “irregular” to differentiate it from a “regular” lattice, where areas are of similar size and shape and are located in a simple pattern, such as squares on a chessboard. It is less common to see spatial econometric analyses focusing on point or line (network) objects or fields. For each unit of observation in an econometric analysis, data on a set of attributes (or variables) are assembled to facilitate the analysis. A digital vector file for storing geometric location information and the associated attribute information is often called a shapefile. While not an absolute requirement, most applied econometric studies unfold through the analysis of data in a shape-file. Technically, this particular term more properly refers only to the digital location and attribute files used by ArcGIS and other GIS software produced and distributed by Environmental Systems Research Institute, Inc. (ESRI). Nevertheless, the term shapefile is becoming a generic term used widely for many GIS.

...