The objective of this study was to compare maternal and midwifery manpower effects of policies for induction of labour (IOL) postdates, using a retrospective cohort design, in a level two maternity unit in a district hospital in South-West England. Primary outcome measures included mode of delivery, admission-delivery interval, midwifery manpower use. Group I consisted of 124 women who underwent IOL at 40+10. Group II were 104 women who underwent IOL at 42 weeks' gestation and 123 women who laboured spontaneously between 40+10 and 42 weeks' gestation. The nulliparous women had a shorter admission-delivery interval when induction was planned for 42 weeks, compared with 40+10 (p = 0.003), and required less frequent use of syntocinon (p = 0.04) and of continuous fetal monitoring (p = 0.02). The caesarean rate was higher in Group I than in Group II (p = 0.04) for nulliparous women only. The earlier induction policy was associated with a higher midwifery manpower requirement for nulliparae (p = 0.002). For parous women, the only difference was the greater use of oxytocin in labour. There was no difference between the groups in duration of labour, analgesia, Apgar scores, admission to neonatal care and meconium aspiration. In conclusion, delaying planned induction by three days was associated with lower medicalisation of labour and manpower needs for nulliparous women.