Non-pharmaceutical interventions, such as social distancing and lockdowns, remain the most viable measure to control the COVID-19 pandemic. Localized lockdowns in small geographic areas have become an important policy intervention to prevent viral spread in cases of resurgence. These localized lockdowns can result in lower social and economic costs compared to larger-scale suppression strategies. In this paper, we adopt a novel approach from the causal inference literature and estimate the effect of localized lockdowns using augmented synthetic controls. Utilizing a rich integrated dataset from Chile, we estimated the direct and indirect causal effects of localized lockdowns on SARS-CoV-2 transmission. Our results show that the effects of localized lockdowns are strongly modulated by their duration and are affected by indirect effects from neighboring geographic areas that are not under lockdown. Our estimations suggest that extending localized lockdowns will slow down the epidemic. However, by themselves, localized lockdowns will be insufficient to control epidemic growth due to indirect effects from neighboring areas unless those contiguous areas also implement lockdowns. These results provide critical empirical evidence about the effectiveness of localized lockdowns in interconnected geographic areas.
