Structural racism in police contact is an important driver of health inequities among the U.S. urban population. Hyper-policing and police violence in marginalized communities have risen to the top of the national policy agenda, particularly since protests in 2020. How did pandemic conditions impact policing? We assess neighborhood racial disparities in arrests after COVID-19 stay-at-home orders in Boston, Charleston, Pittsburgh, and San Francisco census tracts (January 2019–August 2020). Using interrupted time series models with census tract fixed effects, we report arrest rates across tract racial and ethnic compositions. In the weeks following stay-at-home orders, overall arrest rates were 39% lower (95% CI: 37–41%) on average compared to rates the year prior. Although arrest rates steadily increased thereafter, most tracts did not reach pre-pandemic arrest levels. However, despite declines in nearly all census tracts, the magnitude of racial inequities in arrests remained unchanged. During the initial weeks of the pandemic, arrest rates declined significantly in areas with higher Black populations, but average rates in Black neighborhoods remained higher than pre-pandemic arrest rates in White neighborhoods. These findings support urban policy reforms that reconsider police capacity and presence, particularly as a mechanism for enforcing public health ordinances and reducing racial disparities.Supplementary InformationThe online version contains supplementary material available at 10.1007/s11524-021-00598-z.