This paper evaluates the impacts of a public program that introduced access to part-time childcare centers for children younger than four years of age in poor urban areas in Nicaragua. We explore the effects of this program on several measures of children's and parental outcomes. Our identification strategy exploits the original randomization and the distance to the centers, using Instrumental Variables (IV) and Marginal Treatment Effects (MTE) methods to tackle imperfect compliance with the original treatment assignments. We present a theoretical model to rationalize our IV assumptions. We find a positive impact of 0.35 standard deviations on the personal-social domain of a widely used development test, and an impact of 14 percentage points on mothers' work participation. Our results are robust to different econometric specifications. We also find suggestive evidence that quality greatly matters for the impacts at the child level, but not at the mother level.