Cointegration in Panel, Spatial and Spatio‐Temporal Models: Some Recent Advances and Applications

ABSTRACT Cointegration was originally introduced to describe long‐term dependence between nonstationary time series. In this paper, we review some recent theory and applications of the cointegration concept in the context of panels of time series, spatial series, and spatio‐temporal series. Such an extension offers a number of challenges. In particular, it is not clear what the best approach is if the null hypothesis of no cointegration is rejected. An important part of the paper is a brief section on cointegration in climate models. Applications here have given rise to the new discipline of climate econometrics, and we seek to point out some problems pertaining to panels and cointegration.