In this paper we present new evidence on the euro effect on trade. We use a data set containing all bilateral combinations in a panel of 26 OECD countries during the period 1967-2008.From a methodological point of view, we implement a new generation of tests that allow solving some of the problems derived from the non-stationary nature of the data. To this aim we apply panel tests that account for the presence of cross-section dependence as well as discontinuities in the non-stationary panel data. We test for cointegration between the variables using panel cointegration tests, especially the ones proposed by Banerjee and Carrión-i-Silvestre (2010). We also efficiently estimate the long-run relationships using the CUP-BC and CUP-FM estimators proposed in Bai et al. (2009). We argue that, after controlling for cross-section dependence and deterministic trends and breaks in trade integration, the euro appears to generate lower trade effects than predicted in previous studies. JEL classification numbers: C12, C22, F15, F10.