Young children are thought to be vulnerable to separation from the primary caregiver/s. This raises concern about whether early child care enrollment may harm children's development. We use child care assignment lotteries to estimate the effect of child care starting age on early cognitive achievement in Oslo, Norway. Getting a lottery offer lowers starting age by about four months, from a mean of about 19 months in the control group. Lottery estimates show significant score gains for children at age seven. Survey evidence and an increase in labor supply of both mothers and fathers following the offer, suggest that parental care is the most relevant alternative mode of care. We document that the assignment lottery generates balance in observable characteristics, supporting our empirical approach.