The GM estimator of multiple equations panel data model with spatial auto-regressive disturbances

点击翻译按钮获取中文摘要

Panel data is a data form that gained by repeated observation of same individuals, firms or regions. When we use panel data for regression analysis, we have to care about spatial correlation among the disturbances. We proposed the GM estimator for the panel data SUR model with spatial auto-regressive disturbances, and the GM estimator for the panel data model with serial and spatial autoregressive disturbances. These GM estimators need less computational load and require fewer assumptions on distribution of errors than the conventional ML estimators do. Then we examined their performances by numerical experiments.