Exploiting Cross Section Variation for Unit Root Inference in Dynamic Data (Now published in Economics Letters 44 (1), 1994, pp.9-19.)
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(N-?) 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.