Although ecological assemblages frequently depart from neutral model predictions, these discrepancies have not been unambiguously attributed to neutral theory’s core assumption: that community structure is primarily the result of chance variation in birth, death, speciation, and dispersal, rather than the manifestation of demographic differences among species. Using coral communities in Barbados from four time periods during the Pleistocene, we demonstrate that the neutral theory cannot explain coral community similarity distributions, species’ regional abundance distributions, or their local occupancy. Furthermore, discrepancies between the neutral theory and the observed communities can be attributed to violation of the core assumption of species equivalence. In particular, species’ variable growth rates are driving departures from neutral predictions. Our results reinforce an understanding of reef coral community assembly that invokes trade-offs in species’ demographic strategies. The results further suggest that conservation management actions will fail if they are based on the neutral assumption that different coral species are equally able to create live coral cover in the shallow-water reef environment. These findings highlight the importance of developing biodiversity theory that can parsimoniously incorporate species differences in coral reef communities, rather than further elaborating neutral theory.