Abstract
COVID-19 incidence, mobility, and social distancing policies are interdependently linked, with the pathways and level of influence varying over time by region. Using weekly, district-specific time series data, we examined whether the effect and underlying pathways of social distancing policies on changes in COVID-19 incidence and mobility differ between commercial and non-commercial districts in Seoul. We employed the Vector Autoregression with Exogenous variables (VARX) model to analyze the interdependencies among COVID-19 incidences, mobility patterns, and social distancing policies over time from June 2020 to November 2021 across Seoul, including district-specific effects. The model incorporated vaccination coverage as an external variable to account for population-level immunity. Impulse Response Functions (IRFs) were used to interpret the results of the citywide VARX model, illustrating how changes in one variable influenced others over time. Social distancing policies had minimal direct effects on incidence rates but indirectly reduced them by decreasing mobility. Mobility observed two weeks prior demonstrated a positive correlation with current COVID-19 incidence, particularly in commercial districts, indicating that increased movement significantly contributes to transmission. On the other hand, rising COVID-19 cases led to decreased mobility, especially in non-commercial districts, likely driven by fear of infection and heightened self-restriction. Policies were more effective in reducing mobility within commercial districts, likely due to stricter enforcement and higher population density, which amplified compliance impacts. Reduced mobility in commercial districts following stricter policies indicates that targeted restrictions in high-traffic areas can effectively help control COVID-19 transmission in Seoul. In contrast, the more pronounced voluntary movement reductions in non-commercial areas, driven by increased incidence, suggest that enhanced regular testing could play a more significant role in disease control in these regions. While policies reduce mobility, their slower impact on incidence rates, at least a two-week time lag, highlights the need for proactive implementation to achieve measurable decreases in transmission.