Private returns to education versus education spill-over effectsOr what co-workers account for!

In this study we test whether education spill-over effects biase private returns to education. We newly investigate for such effects within narrowly defined occupational groups of a given firm using, going thus a step further than earlier studies analysing such effects at employer level or within regions. Using panel data from a Swiss employer-based labour force survey of 1996 we find clear empirical evidence for education spill-over effects on individual wages (2%). This result proved to be robust against an alternative explanation based on imperfect substitution between highly and less educated workers. Consequently, conventionally calculated private returns to education are shown to be biased upwards.