This paper studies estimation of panel cointegration models withcross-sectional dependence generated by unobserved global stochastic trends.The standard least squares estimator is, in general, inconsistent owing to thespuriousness induced by the unobservable I(1) trends. We propose two iterativeprocedures that jointly estimate the slope parameters and the stochastictrends. The resulting estimators are referred to respectively as CupBC(continuously-updated and bias-corrected) and the CupFM (continuously-updatedand fully-modified) estimators. We establish their consistency and derive theirlimiting distributions. Both are asymptotically unbiased and asymptoticallymixed normal and permit inference to be conducted using standard teststatistics. The estimators are also valid when there are mixed stationary andnon-stationary factors, as well as when the factors are all stationary.
展开▼