Spatial Implications of Telecommuting

If the 2020 surge in working from home became permanent, how would the distribution of jobs and residents within and across U.S. cities change? To study this question, we build a quantitative spatial equilibrium model of job and residence choice with commuting frictions between 4,502 sub-metropolitan locations in the contiguous U.S. A novel feature of our model is the heterogeneity of workers in the fraction of time they work on-site: some workers commute daily, some always work at home, while others alternate between working on-site and remotely. In a counterfactual where remote work becomes more common, residents move from central to peripheral areas within cities, and from large coastal to small interior cities, on average. The reallocation of jobs is less monotonic, with increases both in peripheral locations and in the highest-productivity metropolises. Agglomeration externalities from in-person interactions are crucial for welfare effects. If telecommuters keep contributing to productivity as if they worked on-site, better job market access drives considerable welfare gains, even for those who continue to commute. But if productivity declines in response to the reduction in face-to-face interactions, wages fall and most workers are worse off