Why Working from Home Will Stick

Work from home (WFH) has stabilized at about 25% of all paid workdays as of 2025, 3.5 times its 2019 level. To understand this shift, we collect and analyze over 250,000 survey responses from US workers since 2020. To assess its economic consequences, we build a quantitative equilibrium model that endogenizes work arrangements, workhours, and productivity. By designing our survey and model jointly, we capture key choices and tradeoffs related to WFH. According to our calibrated model, returning to pre-pandemic work arrangements would lower output per hour (inclusive of commute time) by 3.5% and reduce worker welfare by 2.6% on average, with large differences across workers. These results explain the persistence of WFH while raising a broader question: why didn’t the shift happen before 2020? Our answer is that the pandemic compelled workers and firms to experiment with WFH at scale, yielding new information and a re-optimization of work arrangements. We estimate that pandemic-induced experimentation can account for over 90% of the large rise in WFH from 2019 to the post-pandemic period, when the economy settled into a new equilibrium with higher welfare.