Objective To estimate the effect of low maternal age on late fetal death and infant mortality and to estimate the extent of any increase in infant mortality attributable to higher rates of preterm birth among teenagers. Design Population‐based cohort study. Setting Births recorded in the nationwide Swedish Medical Birth Registry. Population All single births to nulliparous women aged 13–24 years (n= 320,174) during 1973–1989. Methods Using information recorded in the medical birth registry, linked to a national education register, the effect of low maternal age on adverse outcomes was estimated using logistic regression analysis. Main outcome measures Late fetal death, neonatal and postneonatal mortality and preterm birth. Results Compared with mothers aged 2CL24 years, adjusted risks of neonatal and postneonatal mortality were significantly increased among mothers aged 13–15 years (odds ratios = 2.7 and 2.6, respectively) and among those aged 16–17 years (odds ratios = 1.4 and 2.0, respectively), while mothers aged 18–19 years had a significant increase in risk of postneonatal mortality only (odds ratio = 1.4). Rates of very preterm birth (≤ 32 weeks), according to maternal age, were: 13–15 years, 5.9%; 16–17 years, 2.5%; 18–19 years, 1.7%; and 20–24 years, 1.1%. The high rates of very preterm birth among young teenagers almost entirely explained the increased risk of neonatal mortality in this group. Conclusions The increased risks of neonatal and postneonatal mortality among young teenagers may be related to biological immaturity. The increase in risk of neonatal mortality is largely explained by increased rates of very preterm birth.