This paper estimates a family fixed-effects model to test whether parental educational investments are reinforcing or compensating differences in child height and body mass index (BMI) within Oportunidades households. The results indicate that allocations are not made in response to endowment differences in isolation, but depend on the interaction of child endowment, gender, and beneficiary status. Moreover, investments are made so as to counterbalance efficiency and equity considerations: widening differences in height but closing gaps in BMI, where both strategies maximize returns given the earnings–endowment function for height and BMI in Mexico. Finally, household socio-economic status matters. Thus, discriminatory responses within the better-off urban beneficiary households are smaller. In contrast, indigenous beneficiary households, which have the lowest average endowment z-scores and fraction spent on education, see the largest differences in educational outlay between child subgroups. While the economic effects seem small, these are measured relative to the mean of the base category. In some cases, then, parents are dedicating nearly a quarter of all expenditures to education for a standard deviation increase in endowment over mean sibling endowment. Hence, it seems that the success of the Oportunidades program in improving overall human capital formation may be masking rising inequalities within recipient households.