After more than 20 years of European Monetary Union (EMU), surprisingly few scientific studies exist which study the growth effects of introducing a common currency in large parts of the European Union. We do so using a large panel (NUTS3 data) of regional data for the EU-15 (GDP plusstructural characteristics of regional units). The introduction of the Euro lends itself naturally to a synthetic control approach as explored e. g. by Abadie, Diamond and Hainmueller (ADH, 2010). We estimate the causal effects of EMU both with the standard ADH-methodology and with a novel approach which is based on minimizing the post-treatment mean squared synthetic control error. While the former may suffer from lack of identification of the predictor weighting matrix V and from inconsistencies in the usage of pre-treatment outcomes, the latter yields consistent estimates and requires not more than ordinary least squares. The results from both approaches differ substantially for some of the regions, but the broad conclusion is the same: EMU has had a sizable detrimental effect on most Eurozone regions with income losses often exceeding 10% of per-capita income over twenty years.