期刊名称：Discussion Papers in Economics / Department of Economics, University of York
出版年度：2015
卷号：2015
出版社：University of York
摘要：We investigate QML estimation of a parametric form for the spatial weight matrix, W, appearing in the mixed regressive, spatial autoregressive (MR-SAR) model and extend the identifiability, consistency, and asymptotic Normality results given by Lee (2004, 2007) to the case when W depends on an unknown parameter, y, that is to be estimated from a single cross-section. Numerical experiments illustrate that the QML estimator works quite well inmoderate sized samples, yielding well-behaved parameter estimates and t-statistics with approximately correct size in most cases. These findings should open the door to a much more flexible approach to the construction of spatial regression models. Finally, the QML estimator using two types of sub-models for the spatial weights is applied to the cross-sectional dataset used in Ertur and Koch (2007), to illustrate the utility of the approach.
Loading...
