Hassler, UweMarmol, FrancescUniversidad Carlos III de Madrid. Departamento de Estadística2010-12-132010-12-131998-01https://hdl.handle.net/10016/9794We consider regressions of nonstationary fractionally integrated variables dominated by linear time trends. The regression errors are short memory, long memory or even nonstationary, and hence allow for a very flexible cointegration model. In case of simple regressions, least squares estimation gives rise to limiting normal distribucions independently of the order of integration of the regressor, whereas the customary t-statistics diverge. We also investigate the possibility of testing for mean reverting equilibrium deviations by means of a residual-based log-periodogram regression. Asymptotic results become more complicated in the multivariate case.application/octet-streamapplication/octet-streamapplication/pdfengAtribución-NoComercial-SinDerivadas 3.0 EspañaNonstationary regressorsStationary or nonstationary errorsLimiting normalityResidual-based cointegration testingworking paperC15C22C32Estadísticaopen access