Abstract for: Population-Level Dynamics of COVID-19 Vaccine Uptake in the US

COVID-19 vaccination uptake varied substantially across U.S.\ states, raising questions about the behavioral and structural mechanisms shaping these patterns. While many epidemiological models treat vaccination behavior as exogenous, real-world uptake emerges from feedback between risk perception, social influence, eligibility, and capacity constraints. Understanding these endogenous processes is essential for explaining observed vaccination trajectories. We develop a simple system dynamics simulation model that captures population-level feedbacks governing vaccination uptake, including social influence, perceived infection risk, heterogeneous willingness, vaccine availability, and capacity limits. Behavioral parameters are estimated using an ODE-informed Tobit regression framework applied to weekly vaccination data from all 50 U.S.\ states and the District of Columbia during 2021--2022, providing an empirical foundation for model calibration. The calibrated model reproduces observed regional vaccination trajectories and reveals two distinct phases of uptake. Early dynamics are governed primarily by responses to vaccine eligibility and capacity constraints, while later dynamics are driven by behavioral responses to changing COVID-19 incidence. Rising case numbers stimulate additional vaccination, whereas cumulative vaccination generally reduces subsequent uptake due to depletion of more willing individuals. Our findings highlight the importance of endogenizing vaccination behavior in epidemic models. Regional heterogeneity in responsiveness to infection risk and baseline vaccination propensity suggests the potential value of moving beyond uniform policy assumptions to better explain observed heterogeneities in vaccine uptake.