New method reduces bias in dynamic panel regressions, improving precision.
The article addresses bias issues in dynamic panel regressions by developing a method that considers cross section dependence. The new estimators reduce bias and improve precision by accounting for error correlation. The study also introduces tests for coefficient homogeneity and proposes a modified test for homogeneous unit roots. An orthogonalization procedure is developed to remove cross section dependence and allow for conventional unit root tests with panel data. Simulations show that the estimation and test procedures perform well in finite samples.
