Economic complexity indicators provide a better understanding of the long-term determinants of countries’ economic growth. However, they do not account for intra-national spatial heterogeneity, even when regions, and not only countries, differ tremendously in their productivity. This paper tries to fill this gap by developing a methodology to estimate economic complexity at the subnational level using a unique trade database on intra-national trade flows for Spain. Specifically, we calculate international and intra-national economic complexity indicators for the provinces of Spain (NUTS-3) for the period running from 1995 to 2016. We find that regions tend to differ in complexity not only in time but also in their international and intra-national trade exposures. We also show that indicators that incorporate both types of trade flows are better predictors of future GDP growth. These findings shed light on the determinants of regional convergence patterns.