In this paper we investigate the causal effect of immigration on trade flows, using Italian panel data covering very small geographical units (NUTS-3). Exploiting the very favorable setup offered by Italy's features – the very high number of countries of origin of immigrants ('super-diversity'), the high heterogeneity of social and economic characteristics of Italian provinces coupled with the absence of cultural (e.g. language) or historical (colonial ties) attractors for immigration –, controlling for a wide set of fixed effects, and dealing with the possible distortions generated by the inappropriate choice of the areal unit, we find a positive association between immigrants' stocks and both export and import flows, in line with the past literature. However, using instrumental variables based on migration enclaves, we show that immigrants have a positive and significant effect on imports only.