Objective: To estimate the impact of caesarean delivery on the incidence of selected neonatal outcomes. Patients and methods: A 15-year, population-based, cohort study (1988–2002) using the Nova Scotia Atlee Perinatal Database compared neonatal outcomes in term newborns born by spontaneous and assisted vaginal delivery, with newborns born by caesarean delivery, with and without labour, using multiple logistic regression. Results: From a total of 142 929 deliveries, there were 27 263 caesarean deliveries, 61% of which were performed in labour. Relative risks were adjusted for year of birth, maternal age, parity, smoking, maternal weight at delivery, hypertensive diseases, diabetes, previous caesarean delivery, use of regional anaesthesia, induction of labour, gestational age at delivery and large and small for gestational age, where significant. Caesarean delivery in labour, but not caesarean delivery without labour, had increased risks for depression at birth and neonatal respiratory conditions compared with spontaneous or assisted vaginal delivery. Compared with spontaneous vaginal delivery and assisted vaginal delivery, the risk of major neonatal birth trauma was decreased for infants after caesarean delivery with labour (odds ratio (OR) = 0.34, 95% CI 0.21 to 0.56 and OR = 0.07, 95% CI 0.04 to 0.11, respectively) and caesarean delivery without labour (OR = 0.20, 95% CI 0.08 to 0.52 and OR = 0.04, 95% CI 0.02 to 0.10, respectively). Conclusion: Caesarean delivery in labour, compared with vaginal delivery, is more likely to be associated with an increased risk for respiratory conditions and depression at birth than caesarean delivery without labour. Caesarean delivery appears protective against neonatal birth trauma, especially when performed without labour.