Although matings between relatives can have negative effects on offspring fitness, apparent inbreeding preference has been reported in a growing number of systems, including those with documented inbreeding depression. Here, we examined evidence for inbreeding depression and inbreeding preference in two populations (Clinton, New York and Davis, California, USA) of the cooperatively breeding American crow (Corvus brachyrhynchos). We then compared observed inbreeding strategies with theoretical expectations for optimal, adaptive levels of inbreeding, given the inclusive fitness benefits and population-specific magnitude of inbreeding depression. We found that low heterozygosity at a panel of 33 microsatellite markers was associated with low survival probability (fledging success) and low white blood cell counts among offspring in both populations. Despite these costs, our data were more consistent with inbreeding preference than avoidance: the observed heterozygosity among 396 sampled crow offspring was significantly lower than expected if local adults were mating by random chance. This pattern was consistent across a range of spatial scales in both populations. Theoretically adaptive levels of inbreeding, given the magnitude of inbreeding depression, were predicted to be very low in the California population, whereas complete disassociative mating was predicted in the New York population. Sexual conflict might have contributed to the apparent absence of inbreeding avoidance in crows. These data add to an increasing number of examples of an “inbreeding paradox,” where inbreeding appears to be preferred despite inbreeding depression.