Spatial Analysis: A Helicopter Tour (From the Beginning to the Frontier)

Abstract

We will begin at the beginning of spatial analysis, around mid-18th century. The introduction will be at a general level of epidemiology. Then we will slowly move to spatial statistics, spatial econometrics and finally to, spatial urban economics. We will touch on the contrast between time-series and spatial dependence, and the need for moving from linear to non-linear models; from fixed to space varying coefficient models. More specifically, in the context of urban econometrics, we will study the price of housing; we will argue that not only the spatial dependence of house prices, but also the dependence in the variability (risk) of prices need to be considered, leading to non-linear spatial autoregressive conditional heteroskedastic (SARCH) model. The major highlight of the talk will be how to test various spatial models, particularly, in the context of possible misspecification.