and the University of Michigan
University of Illinois, Urbana-Champaign
Spatial regression analysis or spatial econometrics
is the collection of statistical and econometric methods
specifically geared at dealing with problems of spatial
dependence and spatial heterogeneity encountered in
cross-sectional (and panel) data sets. The use of spatial
econometric techniques is increasingly common in empirical
work in the social sciences, including urban and regional
economics, criminology, demography and electoral analyses.
The main objective of the course is to review the state
of the art of the specification, estimation and testing
of models that incorporate spatial dependence (and spatial
heterogeneity). While the focus will be on
spatial aspects, the types of methods covered have general validity
in applied statistical work.
The course will include topics such as the specification
of dependent stochastic processes (specifically, various
types of spatial autoregressive models), maximum likelihood
estimation of dependent processes, instrumental variables
and general method of moments estimation, specification
tests, and asymptotic and finite sample properties.
While most of the material will be applied to the standard
regression model, some attention will be paid to panel
data contexts (space-time models) as well as to spatial
probit models. An important aspect of the
course is the application of the spatial regression
techniques in empirical practice, using the SpaceStat software package or general purpose
Prerequisites include intermediate
regression analysis (or intermediate econometrics)
as well as familiarity with spatial data analysis at
the level of ICPSR's "Introduction
to Spatial Data Analysis" course.