Testing for shifts in a time trend panel data model with serially correlated error component disturbances

This paper studies testing of shifts in a time trend panel data model with serially correlated error component disturbances, without any prior knowledge of whether the error term is stationary or nonstationary. This is done in case the shift is known as well as unknown. Following the time series literature, we propose a Wald type test statistic that uses a fixed effects feasible generalized least squares (FE-FGLS) estimator. The proposed test has a chi-square limiting distribution and is valid for both I(0) and I(1) errors. The finite sample size and power of this Wald test is investigated using Monte Carlo simulations.