Previous studies of the returns to education in terms of entrepreneurial profits, have failed to adequately address the problem of a potential selection bias. This paper estimates the returns to schooling based on data from 3000 enterprises in Malawi, using access to land to control for selection. The results show that selection does not affect returns to education; a year of additional schooling increases entrepreneurial profits by approximately 6 per cent. Selection does, however, affect the impact of other variables on profits. Importantly, the negative impact of health problems on profits increases markedly when taking selection into account. In other words, more funds should be allocated to health activities relative to schooling when these results are taken into account.