The nonmarket returns to women's schooling in developing countries are thought to be consistently high. This paper reviews existing studies, discusses future research directions, and presents policy implications on this topic. The review of existing studies suggests that standard studies tend to misrepresent the impact of women's schooling on nonmarket outcomes because of failures to control for: (1) often observed dimensions of women's endowments (e.g., heights), (2) observed community characteristics (e.g., social services); (3) unobserved individual, household and community characteristics; (4) interactions between women's and men's schooling; (5) intergenerational gender links; and (6) subsample selectivity. There is evidence in one provocative study by Pitt and Rosenzweig that subsample selectivity with regard to fertility causes an underestimate of the impact of women's schooling on child health. The available studies that consider the other five possible problems tend to suggest that standard procedures overstate substantially the impact of women's schooling on nonmarket outcomes. Future research profitably could explore the nature of interactions of women's schooling with other inputs, incorporate the effect of women's schooling quality and not just years of schooling, explore possible externalities, investigate other outcomes such as cognitive achievement, consider risk aversion and imperfect information, study inter and intra household transfers and related bargaining and insurance, show more sensitivity to possible measurement errors, and control for the factors mentioned above. The policy implications of research to date include substantial promise to further research in this area and to increasing female schooling for equity reasons. However there does not to date seem to be persuasive evidence that there are market failures that argue for substantial public subsidies for investments in female schooling for productivity reasons, even though the private returns to such investments may be considerable.