Recent evidence on the respective contributions of institutions and trade to income levels across countries has demonstrated that - once endogeneity is considered - institutional quality clearly dominates the effect of trade. We argue that overall trade is not the most appropriate measure for technology diffusion as a source of productivity growth and propose to focus on imports of research and development (R&D) intensive goods instead. Overall, we confirm previous findings that institutions matter most and that overall trade is not positively associated with per-capita income levels. Yet this does not hold for technology trade, as there is a positive and significant linkage between technology imports and income levels. This outcome is robust to various model specifications, including an instrumental variable approach.