Exploiting cross section variation for unit root inference in dynamic data

This paper considers unit root regressions in data having simultaneously extensive cross-section and time-series variation. The standard least squares estimators in such data structures turn out to have an asymptotic distribution that is neither Op(T−1) Dickey-Fuller, nor Op(N12) normal and asymptotically unbiased. Instead, the estimator turns out to be consistent and asymptotically normal, but has a non-vanishing bias in its asymptotic distribution.