Parameter spaces for stationary DGPs in spatial econometric modelling

Unlike the time series literature the spatial econometric literature has not really dealt with the issue of the parameter space. This paper shows that current parameter space concepts for spatial econometric DGPs are inadequate. It proves that the parameter space proposed by Kelejian and Prucha 2008 can result in nonstationary DGPs, while the parameter space proposed by Lee and Liu 2010 can be too restrictive in applied cases. Furthermore it is discussed that the practice of row standardizing lacks a mathematical foundation. Due to these problems concerning the current parameter space consepts, this paper provides a new de...nition for the spatial econometric parameter space. It is able to show which assumptions are necessary to give row standardizing the needed mathematical foundation. Finally two additional applications for the new parameter space de...nition concerning models with group interaction and panels with fixed cross section sample size are provided. Both applications result in parameter spaces that are substantially larger than the ones the literature would so far considered to be stationary.