Description

This Master Prompt focuses on panel-data settings in which current outcomes depend on their own historical values and where one or more explanatory variables may be endogenous, predetermined, weakly exogenous, or correlated with unobserved effects. It guides researchers in moving systematically from a static panel specification to an explicitly dynamic architecture; identifying theoretically justified lag structures; understanding the econometric consequences of including lagged dependent variables; distinguishing strict exogeneity, predeterminedness, and endogeneity; constructing defensible instrumental variables; formulating valid moment conditions; and selecting among difference GMM, system GMM, IV, bias-corrected method-of-moments, bias-corrected within-group, profile GMM, QML, and related dynamic-panel estimators.