The article uses continuous wavelet analysis for the study of certain export-import time series. It is a novel methodology, with the promise, inter alia, to detect features of processes that may remain hidden for traditional methodologies. The application of wavelets has been rapidly increasing in economics and finance, but to our best knowledge, it has never been used to analyse trade data. In this study export-import series of 10 EU economies and the EU28 are studied. Our results bear on trade forecasting: it seems that the 2008 crisis increased the cohesion of trade in the short term only temporarily, and one can expect significant co-movement only at scales of 3 years and above. Also, we suggest a new classification of small EU economies with respect to integration, inducing interesting new research questions.