Most studies on gender gaps in agricultural productivity leverage within-household differences between plots managed by women and men. Such a gender-based division of plot management simplifies empirical tests for productivity differences, but it is not a common arrangement for agricultural households outside some locations in sub-Saharan Africa. In most rural households, women and men jointly participate in production, which complicates identification of gender- based productivity differences. This study proposes a broader empirical test of productivity gaps that applies to such systems, and that is rooted not explicitly in gender but in gender-based inequities. Specifically, we explore productivity gaps in rice-cultivating Indian households, where women and men perform specific and distinct cultivation tasks. We measure productivity gaps based on the differential use of family and hired female labor across households, then compare them with gaps based on the differential use of family and hired male labor. Using plot-level data, we identify significant gender-based productivity gaps after controlling for input use, plot- and household-level characteristics, and using village fixed effects and machine learning estimators to address selection and model misspecification concerns. Based on this identification strategy, households using family female labor have lower agricultural productivity, on average, than those also hiring female workers, such that foregone production value is greater than the cost of hiring women. We find suggestive evidence that this gap stems from skill differences between hired and family female workers. In contrast, we find no evidence of a similar gap based on the differential use of family and hired male labor. Overall, household welfare is lower because of gender inequities that shape women's work opportunities. These findings highlight the potential productivity implications of expanding women's labor choices, including both on- and off-farm job opportunities.