It is well known that age at arrival determines immigrant childbearing, which is often seen as evidence of adaptation. However, the mechanisms that link age at arrival and fertility remain unclear and have yet to be examined empirically. In response, we carry out a case study of Sweden, using longitudinal register data for the whole population. More specifically, we focus on immigrants who arrived as children and use the parametric g‐formula to examine whether the effect of age at arrival on fertility is mediated by time‐varying socio‐economic processes, including education, employment and citizenship. Our results show that in general these mechanisms play a relatively minor role in mediating the effect of age at arrival. The direct effect of age at arrival is much more substantial, thereby highlighting the importance of other mediating processes that are unobserved in our data, such as language acquisition or the development of social networks.