We consider within-group estimation of higher-order autoregressive panel models with exogenous regressors and fixed effects, where the lag order is possibly misspecified. Even when disregarding the misspecification bias, the fixed-effect bias formula is quite different from the correctly...

This paper considers higher-order autoregressive (AR(p)) panel models with fixed effects, where the lag order p is unknown and possibly misspecified. A pooled least squares estimator is considered and its asymptotic biases are studied. Specifically, we first extend the N-asymptotic bias formula...

Panel unit root tests represent a significant advancement in addressing the low power of unit root tests by exploiting cross-sectional and time-series information. In this article we employ Monte Carlo techniques to quantify the power improvements due to cross-sectional information and assess...