The evolution of social behaviour from solitary antecedents has traditionally been attributed to inclusive fitness benefits. Direct fitness components account for individuals' lifetime reproductive success through the production of offspring. In contrast, indirect fitness components account for the transmission of genes via aiding relatives and may explain apparent altruism among some members of social groups. Hamilton's rule gives the conditions under which altruism may evolve, but measuring the parameters in Hamilton's rule has proved to be very difficult for most social species. Here we quantify direct and indirect fitness in a facultatively social bee, Ceratina australensis, capable of both solitary and social nesting. Social colonies of this species contain one reproductive (the primary) and one nonreproductive female (the secondary), and it is therefore straightforward to measure the lifetime fitness of each individual. In the absence of the primary, the secondary female is fully capable of reproduction and produces equivalent numbers of offspring as solitary females. Per capita brood production was lower in social colonies than solitary nests. We found that inclusive fitness arguments do not explain social nesting in this bee. Moreover, social nesting cannot be attributed to subfertility or manipulation. The frequency of natal nest reuse is highly correlated with the frequency of social nesting in this and other Ceratina species. In other words, social nesting might be influenced by dispersal patterns and latent genetic differences in any tendency for natal philopatry. We argue that for C.australensis and perhaps many species with simple forms of sociality, life history traits such as nest reuse and natal philopatry are key selective factors for the origin of group living. Future work on this and other Ceratina species will reveal genetic, life history and ecological correlates with transitions in social evolution and their potential causes.