On the Large-Sample Estimation of Regression Models with Spatial- Or Network-Effects Terms: A Two-Stage Least Squares Approach

In this paper we apply the two-stage leastsquares (2SLS) estimation technique to produce regression estimators for linear models that incorporate spatialor network-effects terms. Such models are often necessary to adequately represent social processes in data that have been aggregated for politically or administratively defined areas such as census tracts, cities, counties, or states. They are also necessary to represent data from individuals or organizations connected in explicit social networks. We review maximum likelihood (ML) estimators of the spatial-