Informed management of Ethiopian wolf Canis simensis populations necessitates an understanding of their dynamics and of the social groups that constitute them. Benefitting from a natural experiment occasioned by a disease outbreak, we studied the demography and spatial behaviour of 14 Ethiopian wolf packs across 16 population years in the Bale Mountains, testing for density-dependent regulation and comparing high-density versus recovery phases. The rate of population growth was density-dependent across the study period. As for many other long-lived species with delayed reproduction, adult survival (to 2 years or more, when wolves become adults) contributed most to population change, followed by pup survival (to 1 year), the latter with a steeper response to density change. The number of pups per pack (= all pups divided by the number of packs) also affected population growth. As typically only the dominant female in the pack breeds, the number of pups born each year depended on the number of packs in the population and not on population density. Packs produced larger litters during the recovery from the disease outbreak, but delays in the formation of new social groups created a reproductive deficit, and the rate of population growth at low density was lower than expected by density dependence. Wolf packs that survived the epizootics remained cohesive and expanded their territory. Only once they had attained large membership, one pack split up and others provided dispersers that coalesced into new packs. With this strategy, the surviving wolves deterred colonization by outsiders and eventually inherited part of their natal range. We discuss the implications of our findings for the future management of Ethiopian wolf populations and in the context of the population dynamics of other species with similar life-history traits.