Two-stage generalized moment method approach for bidimensional random coefficient autoregressive models.

A two-dimensionally indexed random coefficients autoregressive models (2D−RCAR) and the corresponding statistical inference are important tools for the analysis of spatial lattice data. The study of such models is motivated by their second-order properties that are similar to those of 2D− (G)ARCHwhich play an important role in spatial econometrics. In this article, we study the asymptotic properties of two-stage generalized moment method (2S−GMM) under general asymptotic framework for 2D−RCAmodels. So, the efficiency, strong consistency, the asymptotic normality, and hypothesis tests of 2S−GMMestimation are derived. A simulation experiment is presented to highlight the theoretical results. [ABSTRACT FROM PUBLISHER]