We study the effects of population decline on the spatial distribution of economic activity. Empirically, we document that population growth is negatively associated with changes in spatial concentration, both across and within countries. We then show theoretically that an aggregate population decline raises spatial concentration when housing supply elasticities decrease or expenditure shares increase with population density across regions, both of which are constant in benchmark spatial models. To quantify this mechanism, we estimate a dynamic spatial model using Japanese data and regional variation in housing supply elasticities and expenditure shares. In response to the projected population decline, the model predicts that Tokyo’s population share rises from 10% to 40%, while low-density regions empty out. A social planner would like to increase concentration further by making transfers from low- to high-density regions.