A1046
Title: An estimator for dynamic linear panel data models based on nonlinear moment conditions
Authors: Andrew Adrian Yu Pua - De La Salle University Manila (Philippines)
Markus Fritsch - University of Passau (Germany)
Joachim Schnurbus - University of Passau (Germany) [presenting]
Abstract: An instrumental variables (IV) estimator based on nonlinear (in parameters) moment conditions that may resolve identification problems regarding the lag parameter when estimating linear dynamic panel data models is proposed. The estimator is applicable in the unit root, near-unit root, and non-unit root case. Consistency and asymptotic normality of the estimator are shown when the cross-section dimension is large and the time series dimension is either fixed or large, and the improvement in the rate of convergence compared to existing estimators is discussed. While the estimator point identifies the lag parameter when the lag parameter is one, it yields two distinct solutions otherwise. A selection rule is proposed and analyzed for the latter case, which is supposed to identify the consistent root asymptotically.
