This paper addresses the estimation of peer group effects on a fertility decision. The peer group is composed of neighbors with similar socio-demographic characteristics. In order to deal with the endogeneity problem associated to the estimation of neighborhood effects, an instrumental variables procedure is performed. To control for the reáection problem, usual in linear e§ects models, this paper uses an identification strategy that relies on the definition of peer groups at the individual level. This paper provides evidence that peer e§ects explain the age at which poor women in MedellÌn (Colombia) decide to have their firstborn. These social forces are hazardous factors that may increase the incidence of adolescent pregnancy.