For marine invertebrates such as corals, restricted dispersal of sperm and/or larvae have been invoked to explain large heterozygote deficits and population subdivision apparent in many genetic surveys. Equally though, for the many corals and other invertebrates that are hermaphroditic, inbreeding through self-fertilisation may also account for heterozygote deficits. Flexibility of mating systems to allow at least some level of self-fertilisation may be favoured by selection, as this would facilitate the founding of new populations by low numbers or densities of colonists. While tests for self-compatibility are relatively easy for broadcast-spawning corals, experimentally determining the level of selfing in corals that have internal fertilisation is near impossible; ironically, it is these brooding species that are considered good colonists and hence the most likely to display a tolerance for self-fertilisation. Here we used allozyme data to provide the first rigorous estimate of outcrossing rates in a brooding coral Acropora palifera, and compare these values with indirect estimates of mating system based on adult genotype frequencies. We found that within each of 2 sites (150 and 300 m2) at Wistari Reef on the southern Great Barrier Reef, estimated outcrossing rates were not significantly less than rates expected from random mating (t = 0.96 ± 0.07 and t = 0.92 ± 0.09). Levels of biparental inbreeding (0.02 and 0.15) and correlated paternity estimates (0.14 and 0.39) were intermediate to low, implying that broods typically had multiple male parents. The adult populations at both sites showed evidence of greater levels of heterozygote deficits than could be explained based on levels of outcrossing estimated from the genotypes of broods, implying that either outcrossing rates vary over time or that population genetic structure is strongly influenced by other factors such as restricted larval dispersal.