We investigate the importance of various mechanisms by which child care policies can affect life-cycle patterns of employment and fertility among women, as well as long-run cognitive outcomes among children. A structural life-cycle model of employment, fertility, and child care use is estimated using Norwegian administrative data. The estimation exploits a large-scale child care reform, which provided generous cash transfers to mothers who did not use formal child care facilities. Combining with administrative data on national test scores, we examine the effects of mother's behavior on long-run cognitive outcomes of children, via estimating a cognitive ability production function that corrects for the endogeneity of inputs. We find that the child care reform generates sizable changes in employment and fertility decisions, especially among low-education women. This leads to lower reading scores among children, primarily as a result of mothers shifting away from formal care and becoming employed. Simulation results suggest that a partial reform, in which workers are ineligible for cash transfers, can generate a more balanced impact on the population. The implications of tax policy and maternity leave are also investigated.