Estimating the Causal Effect of Policy Interventions in the Presence of Spillovers

Public policy interventions are commonly evaluated using the difference-in-differences approach. However, this approach does not directly account for the effect of the policy ``spilling over" to neighboring regions such as nearby cities or states. To address this, we extend difference-in-differences methods to identify the causal effects of policy interventions under various spillover conditions. We propose doubly robust estimators for the average treatment effect on the treated and on the neighboring control. The new estimators relax the standard assumptions on interference and model specification. In addition, we formally define a new causal estimator for the average treatment effect on the treated as a function of neighborhood exposure to the policy intervention. Importantly, our approach allows one to generalize the causal effect of a policy change to other target populations that may be different than the original study population. For instance, the effect of the beverage tax implemented in Philadelphia could potentially be transported to other cities with very different contextual and demographic characteristics.