Title: Matrix spatial specification models
Authors: Samantha Leorato - University of Rome Tor Vergata (Italy) [presenting]
Andrea Martinelli - University of Insubria (Italy)
Abstract: Spatial linear regression models with spatial dependence in the errors and in the dependent variable are studied. The spatial dependence is modeled by arbitrary matrix functions, $V$ and $M$ respectively, indexed by a scalar parameter and, eventually, by two (possibly distinct) weight matrices, $D$ and $W$. This family of models encompasses the main models used in the spatial econometric literature, such as SARAR and MESS models. We define the quasi maximum likelihood estimator and study its asymptotic properties under non-Gaussian errors. By doing so, we provide some insights into the difference between specifications, with emphasis on advantages and shortcomings as well as on interpretation of the parameters and correspondences between models.