Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This paper studies dynamic spatial panel data models with common shocks to deal with both weak and strong cross-sectional correlations. Weak correlations are captured by a spatial structure and strong correlations are captured by a factor structure. The proposed quasi-maximum likelihood estimator (QMLE) is capable of handling both types of cross sectional dependence. We provide a rigorous analysis for the asymptotic theory of the QMLE, demonstrating its desirable properties. Heteroskedasticity is explicitly allowed. This is important because QML is inconsistent in the presence of heteroskedasticity while homoskedasticity is imposed. We further show that when heteroskedasticity is estimated, the limiting variance of QMLE is not a sandwich form regardless of normality. Monte Carlo simulations show that the QMLE has good finite sample properties.