Despite the introduction of childcare subsidies in many countries, the cost of childcare is still thought to hinder parental employment. Many governments are considering increasing the generosity of their childcare subsidies, but the a priori effect of such a policy is ambiguous and little is known empirically about its likely impact. This paper compares the effects on parents' labour supply of offering free part-time childcare and of expanding this offer to the whole school day in England using an empirical strategy which, unlike previous studies, exploits both date of birth discontinuities and panel data. We find that the provision of free part-time childcare has little, if any, causal impact on the labour market outcomes of mothers or fathers. Increasing the number of hours of free childcare to cover a full school day, however, leads to significant increases in the labour supply of mothers whose youngest child is eligible, with impacts emerging immediately and increasing over the months following entitlement.