In a recent contribution, Redding and Schott (2003) add human capital to a two sector NEG model, highlighting that remoteness represents a penalty that gives disincentives to invest in human capital. But is this hypothesis consistent with long-term evidence? We test the persistence of this effect at the regional level in an historical setting. The results show that market access has a significant positive influence on human capital in OLS, Tobit and IV regression models. Thus, the paper confirms the 'penalty of remoteness' hypothesis for Europe in the long run.