cient GMM estimation of spatial dynamic panel data models with xed e ¤ ects

This paper proposes the GMM estimation of the spatial dynamic panel data model with …xed e¤ects when n is large, and T can be large, but small relative to n. By eliminating …xed e¤ects to begin with, we investigate asymptotic properties of the estimators, where exogenous and predetermined variables are used as instruments. For the spatial dynamic panel data model, as compared with the dynamic panel data model, we have not only more linear moment conditions due to spatial e¤ects, but also quadratic moment conditions. We stack up the data and construct the best linear and quadratic moment conditions. An alternative approach is to use separate moment conditions for each period, which gives rise to many moment estimation. We show that these estimators are p nT consistent, asymptotically normal, and can be relatively e¢ cient. We compare these approaches on their …nite sample performance by Monte Carlo.