Data and Code for: Cognitive Hubs and Spatial Redistribution

In the US, cognitive non-routine (CNR) occupations are disproportionately and increasingly represented in large cities. To study the allocation of workers across cities, we propose a quantitative spatial equilibrium model with multiple industries employing CNR and non-CNR workers. Productivity is city-industry-occupation specific and, as we estimate, partly determined by externalities that depend on local occupation shares and total employment. An optimal policy that benefits workers equally, incentivizes the formation of cognitive hubs in large cities. It also creates higher overall activity in small cities, greater industrial specialization in both the largest and smallest cities, and greater diversification in medium-sized cities.<br>