This study tests and compares the two major approaches to the modelling of human capital in growth regressions, i.e. the Lucas and the Nelson-Phelps approach, in the context of developing country models with international knowledge spillovers. On balance, the results seem to favour the Nelson-Phelps approach. Using human capital stock variables instead of flow variables, a positive role for human capital in the absorption of international knowledge spillovers other than embodied R&D spillovers is confirmed. The results suggest the importance of distinguishing between different types of international knowledge spillovers, as well as between different human capital sub-categories.