Existing literature suggests that girls are differentially affected by (1) income shocks and (2) changes in bargaining power. However, these analyses do not shed light on the actual sources of discrimination, i.e. whether differential treatment is the result of maximization given differential returns and opportunity costs or a more deeply entrenched notion of gender bias. In particular, the majority of studies neglect the role of household production and children’s time allocation in gender discrimination. Using data from the 1990 Indonesian Population Census and 1993 Indonesian Socioeconomic Survey, this paper seeks to identify and quantify the potential sources of discrimination, namely preferences, income and time allocation. To identify the distinct sources of discrimination, I utilize three household types, each highlighting different levels of parental involvement and exogenous shocks to income, as well as their interaction. A household fixed effects model is employed to control for unobservable characteristics which may be correlated with household type. Results indicate that a reduction in the amount of time available for household production, in most cases, reduces the probability that daughters will be enrolled in school, relative to their brothers, but a reduction in household income alone may have a positive effect on girls’ school enrollment relative to their male siblings. Increasing mothers’ bargaining power has the largest impact on closing the gender gap.