The striking cross-national variation in socio-economic disparities in skills gaps among 15-year-olds, and the evolution of these gaps between the ages of 15 and 27, raises the question of what policies and institutional arrangements may explain such variability. Extensive policy analysis and research has been devoted to the features of education systems which are most strongly associated with such socio-economic gradients (or the lack of them) in literacy and numeracy. However, much less is known about which factors contribute to narrowing or widening socio-economic gaps after the end of compulsory schooling. Results on the widening gap at the bottom end of the performance distribution identify a target group for policy interventions – socio-economically disadvantaged students who are low-achievers at the age of 15. These results also help to formulate hypotheses as to why gaps widen in many countries after schools are no longer able to exert their equalising effect, since this is the group which is less likely to enjoy opportunities for further skill development through education and training.