Mark Pennington, "The Irrelevance of Educational Externalities":
Seen in this context, there is little reason to suppose that education generates policy relevant external effects. First, it is not clear that benefits generated by private decisions are uncompensated or that they might not readily be internalised via private mechanisms. The fact that an educated person may raise the productivity of others with whom they interact is likely to be reflected in that person commanding a higher wage or salary than a less educated person. One would expect profit-seeking firms to pay more for workers who raise the productivity of their co-workers.
Second, insofar as decisions over education do generate uncompensated effects these seem likely to be infra-marginal and thus policy irrelevant. Given that people have purely private inducements to invest in literacy and numeracy– such as the greater likelihood of finding a job – it is not clear why governmental efforts to encourage additional investment would be worthwhile.