COINTEGRATION VECTOR ESTIMATION BY DOLS FOR A THREE-DIMENSIONAL PANEL

This paper extends the asymptotic results of the dynamic ordinary least squares (DOLS) cointegration vector estimator of Mark and Sul (2003) to a three-dimensional panel. We use a balanced panel of N and M lengths observed over T time periods. The cointegration vector is homogenous across individuals but we allow for individual heterogeneity using different short-run dynamics, individual-specific fixed effects and individual-specific time trends. Both individual effects are considered for the first two dimensions. We also model some degree of cross-sectional dependence using time-specific effects.This paper was motivated by the three-dimensional panel cointegration analysis used to estimate the total factor productivity for Colombian regions and sectors during 1975-2000 by Iregui, Melo and Ram´ırez (2007). They used the methodology proposed by Marrocu, Paci and Pala (2000); however, hypothesis testing is not valid under this technique. The methodology we are currently proposing allows us to estimate the long-run relationship and to construct asymptotically valid test statistics in the 3D-panel context.