"Maternity leaves can affect mothers' and infants' welfare if they first affect the amount of time working women stay at home post birth. We provide new evidence of the labor supply effects of these leaves from an analysis of the introduction and expansion of job-protected maternity leave in Canada. The substantial variation in leave entitlements across mothers by time and space is likely exogenous to their unobserved characteristics. This is important because unobserved heterogeneity correlated with leave entitlement potentially biases many previous studies of this topic. We find that modest mandates of 17-18 weeks do not increase the time mothers spend at home. The physical demands of birth and private arrangements appear to render short mandates redundant. These mandates do, however, decrease the proportion of women quitting their jobs, increase leave taking, and increase the proportion returning to their pre-birth employers. In contrast, we find that expansions of job-protected leaves to lengths up to 70 weeks do increase the time spent at home (as well as leave-taking and job continuity). We also examine whether this increase in time at home affects infant health, finding no evidence of an effect on the incidence of low birth weight or infant mortality"--National Bureau of Economic Research web site.