An Alternative to Partial Regression in Maximum Likelihood Estimation of Spatial Autoregressive Panel Data Model
An Alternative to Partial Regression in Maximum Likelihood Estimation of Spatial Autoregressive Panel Data Model

Summary/Abstract: In this paper an alternative procedure to partial regression is introduced. The presented procedure can be used in maximum likelihood estimation of spatial autoregressive model. Under certain assumptions on the dimension of certain invariant space associated with spatial weight matrix a feasible procedure is formulated, which can be used to handle large class of fixed effect designs. This is done at the expense of possibly decreased number of degrees of freedom in the Gaussian log likelihood function. Additionally, a statement on asymptotic behaviour of presented estimator is given.

• Details
• Contents